special theme: educational technology · educational technologies, “edtech”, illustrating the...

ERCIM NEWSNumber 120 January 2020ercim-news.ercim.eu

Educational Technology

Special theme:

Joint

SPECIAL THEME

The special theme “Educational Technology” has beencoordinated by Vassilis Katsouros (Athena Research andInnovation Centre) and Martin Hachet (Inria)

Introduction to the special theme4 Educational Technology

by Vassilis Katsouros (Athena Research and InnovationCentre) and Martin Hachet (Inria)

6 Α Logic-Based Affective Tutoring System

by Achilles Dougalis and Dimitris Plexousakis (ICS-FORTH)

7 Learning Introductory Programming with Smart

Learning Environment

by Boban Vesin (University of South-Eastern Norway),Katerina Mangaroska and Michail Giannakos (NTNU)

9 Interoperable Education Infrastructures: A

Middleware that Brings Together Adaptive, Social

and Virtual Learning Technologies

by Christopher Krauss and Manfred Hauswirth(Fraunhofer FOKUS)

10 Fraunhofer IAIS IoT Programming Language

NEPO® in the Open Roberta® Lab

by Thorsten Leimbach (Fraunhofer IAIS), Daria Tomala(Fraunhofer IAIS)

12 Modelling Student Learning and Forgetting for

Optimally Scheduling Skill Review

by Benoît Choffin (LRI), Fabrice Popineau (LRI) andYolaine Bourda (LRI)

13 Ontology-based Learning Analytics in Medicine

by Fabrice Jouanot (Univ. Grenoble Alpes), OlivierPalombi (Univ. Grenoble Alpes, CHU de Grenoble) andMarie-Christine Rousset (Univ. Grenoble Alpes, IUF)

15 Blockchain for Education: Lifelong Learning

Passport

by Wolfgang Prinz, Sabine Kolvenbach and RudolfRuland (Fraunhofer FIT)

16 An Educational Platform for Logic-based Reasoning

by Dimitrios Arampatzis (FORTH-ICS), MariaDoulgeraki (FORTH-ICS), Michail Giannoulis (Univ. ofCrete), Evropi Stefanidi (FORTH-ICS), and TheodorePatkos (FORTH-ICS)

17 Authoring Game-Based Learning Activities that are

Manageable by Teachers

by Pedro Cardoso (FEUP/ FBAUP and INESC TEC),Leonel Morgado (Universidade Aberta and INESCTEC), and António Coelho (FEUP and INESC TEC)

19 LudiMoodle: Adaptive Gamification to Improve

Learner Motivation

by Élise Lavoué (Université Jean Moulin Lyon 3)

ERCIM NEWS 120 January 2020

Editorial Information

ERCIM News is the magazine of ERCIM. Published quarterly, it reports

on joint actions of the ERCIM partners, and aims to reflect the contribu-

tion made by ERCIM to the European Community in Information

Technology and Applied Mathematics. Through short articles and news

items, it provides a forum for the exchange of information between the

institutes and also with the wider scientific community. This issue has a

circulation of about 6,000 printed copies and is also available online.

ERCIM News is published by ERCIM EEIG

BP 93, F-06902 Sophia Antipolis Cedex, France

+33 4 9238 5010, [email protected]

Director: Philipp Hoschka, ISSN 0926-4981

Contributions

Contributions should be submitted to the local editor of your country

Copyright�notice

All authors, as identified in each article, retain copyright of their work.

ERCIM News is licensed under a Creative Commons Attribution 4.0

International License (CC-BY).

Advertising

For current advertising rates and conditions, see

https://ercim-news.ercim.eu/ or contact [email protected]

ERCIM�News�online�edition:��ercim-news.ercim.eu/

Next�issue:�

April 2020: The Climate Challenge

Subscription

Subscribe to ERCIM News by sending an email to

[email protected]

Editorial�Board:

Central editor:

Peter Kunz, ERCIM office ([email protected])

Local Editors:

Austria: Erwin Schoitsch ([email protected])

Cyprus: Georgia Kapitsaki ([email protected]

France: Christine Azevedo Coste ([email protected])

Germany: Alexander Nouak ([email protected])

Greece: Lida Harami ([email protected]),

Athanasios Kalogeras ([email protected])

Hungary: Andras Benczur ([email protected])

Italy: Maurice ter Beek ([email protected])

Luxembourg: Thomas Tamisier ([email protected])

Norway: Monica Divitini, ([email protected]),

Are Magnus Bruaset ([email protected])

Poland: Hung Son Nguyen ([email protected])

Portugal: José Borbinha ([email protected])

Sweden: Maria Rudenschöld ([email protected])

Switzerland: Harry Rudin ([email protected])

The Netherlands: Annette Kik ([email protected])

W3C: Marie-Claire Forgue ([email protected])

Cover�illustrtation: source: Shutterstock.

ERCIM NEWS 120 January 2020 33

20 Can Tangible Robots Support Children in Learning

Handwriting?

by Arzu Guneysu Ozgur, Barbara Bruno, ThibaultAsselborn and Pierre Dillenbourg (EPFL)

22 “You Tell, I Do, and We Swap until we Connect All

the Gold Mines!”

by Jauwairia Nasir, Utku Norman, Barbara Bruno andPierre Dillenbourg (EPFL)

23 Gestures in Tangible User Interfaces

by Dimitra Anastasiou and Eric Ras (LuxembourgInstitute of Science and Technology)

25 Kniwwelino: A Microcontroller-based Platform

Introducing Children to Programing and Electronics

by Valérie Maquil and Christian Moll (LuxembourgInstitute of Science and Technology)

26 Virtual Simulation Environment for Medical

Training

by Thomas Luiz, Dieter Lerner, Dominik Schnier(Fraunhofer IESE)

27 Mixed Reality and Wearables in Industrial Training

by Ralf Klamma (RWTH Aachen University), IstvánKoren (RWTH Aachen University) and Matthias Jarke(Fraunhofer FIT and RWTH Aachen University)

29 Personalized Interactive Edutainment in Extended

Reality (XR) Laboratories

by Aris S. Lalos (ISI, ATHENA R.C.), Chairi Kiourt(ILSP, ATHENA R.C.), Dimitrios Kalles (Hellenic OpenUniversity) and Athanasios Kalogeras (ISI, ATHENAR.C.)

30 Music in Education through Technology

by Maximos Kaliakatsos-Papakostas, Kosmas Kritsis,Vassilis Katsouros (Institute for Language and SpeechProcessing, Athena Research and Innovation Centre)

32 Dance Education and Digital Technologies

by Katerina El Raheb and Yannis Ioannidis (“Athena”Research and Innovation Center and National andKapodistrian University of Athens)

33 Intelligent Classroom: Materialising the Vision of

Ambient Intelligence for Education

by Asterios Leonidis, Maria Korozi, Margherita Antonaand Constantine Stephanidis (FORTH-ICS)

35 Innovative Monitoring of Learning Habits and

Motivation in Undergraduate Mathematics

Education

by Brigitta Szilágyi (Budapest University of Technologyand Economics), Szabolcs Berezvai (BudapestUniversity of Technology and Economics) and DanielHorvath (EduBase Online Ltd.)

RESEARCH ANd INNOvATION

41 KANDINSKY Patterns: A Swiss-Knife for the Study

of Explainable AI

by Andreas Holzinger (Medical University Graz andAlberta Machine Intelligence Institute Edmonton,Canada), Peter Kieseberg (University of AppliedSciences, St.Poelten) and Heimo Mülller (MedicalUniversity Graz)

43 Building Reliable Storage Systems

by Ilias Iliadis (IBM Research – Zurich Laboratory)

45 Browse, Visualise, Analyse EU Procurement Data

by Ahmet Soylu and Till C. Lech (SINTEF)

46 More on 5G: Millimetre-Waves

by Gregor Dürrenberger (FSM) and Harry Rudin

ANNOuNCEMENTS

48 FMICS 2020: 25th International Conference on Formal

Methods for Industrial Critical Systems

48 ERCIM Membership

49 ERCIM “Alain Bensoussan” Fellowship Programme

50 Dagstuhl Seminars and Perspectives Workshops

50 Tim Berners-Lee Launches Global Action Plan to Pre-

vent "Digital Dystopia"

50 HORIZON 2020 Project Management

51 W3Cx Introduction to Web Accessibility – New Online

Course

51 HAL is Opening Up to Software

51 David Chaum and Guido van Rossum awarded Dijkstra

Fellowship

36 When a Master of Sciences on EdTech becomes an

International Community

by Margarida Romero, Saint-Clair Lefèvre (UCA,INSPE, LINE) and Thierry Viéville (Inria)

38 DE-TEL - A European initiative for Doctoral

Education in Technology-Enhanced Learning

by Mikhail Fominykh and Ekaterina Prasolova-Førland(NTNU)

39 Ethical Teaching Analytics in a Context-Aware

Classroom: A Manifesto

by Romain Laurent(Univ. Grenoble Alpes/LaRAC),Dominique Vaufreydaz (Univ. Grenoble Alpes, CNRS,Inria, Grenoble INP, LIG), and Philippe Dessus (Univ.Grenoble Alpes/LaRAC)

CONTENTS


Special Theme: Educational Technology

This special theme addresses the state of the art ineducational technologies, “EdTech”, illustratingthe range of scientific fields and challenges facedby the research community when it comes to inte-grating tools and systems that apply to real-lifelearning situations.

Education, either formal or informal, is a keydriver for the future of our societies; it fosters per-sonal fulfilment and development, social inclusionand active citizenship, as well as generating inno-vation and economic activities. Right now, digitaltools are opening up promising new opportunitiesto take education to the next level.

Digital tools in education go far beyond the intro-duction of computers to schools. The introductionof computers to children can even have counter-productive effects on their ability to acquireknowledge and skills. It is vital that we take aholistic approach to educational technologies,where the learner and the teacher stay at the centreof the loop.

As a consequence, modern educational technolo-gies are emerging from multidisciplinary research,building notably on advances in pedagogical andlearning theories, educational psychology, interac-tion technologies and artificial intelligence. Thisspecial issue illustrates the richness of the currentresearch in EdTech, where human factors, softwareand hardware technologies, and even organisa-tional and ethical considerations all contributetowards building tomorrow's education, either informal, informal, or professional learning contexts.

Intelligent tutoring systems that adapt learningcontent to the individual’s progress are becomingmore sophisticated with the ability to consider thelearner’s emotions and learning style, as illustratedby Dougalis and Plexousakis (p. 6). Vesin,Mangaroska and Giannakos (p. 7) present a per-sonalised and adaptive tutoring system covering acomplex interplay of content, tasks, instructions,social dynamics and learning analytics to teachintroductory programming to university students.A common learning middleware in the article by

Krauss and Hauswirth (p. 9) facilitates contentfrom different learning management systems andenriches it with innovative technologies such asgamification, social learning, virtual reality andlearning recommenders. Leimbach and Tomala (p.10) propose an integrated programming environ-ment combined with a graphical programming lan-guage that allows learners to code a wide range ofrobotic systems, promoting STEM education toyoung children.

Choffin, Popineau and Bourda (p. 12) suggestmodelling student learning and forgetting for opti-mally scheduling distributed practice of skills, andJouanot , Palombi and Rousset (p. 13) propose anontology-based method for making learning ana-lytics transparent and explainable using a querylanguage that allows users to express specificneeds of data exploration and analysis. As certifi-cates play an important role in education, and indi-vidual learning records become essential forpeople’s professional careers, the blockchain foreducation platform (Prinz, Kolvenbach andRuland, p. 15) introduces a secure and intuitivesolution for issuing, sharing, and validating educa-tional certificates.

Motivation is an essential lever for engaging stu-dents in learning processes. Hence, severalauthors have explored how motivation can bemaintained and enhanced through gamificationmechanisms. This is notably the case for EduBAI(Arampatzis et al., p. 16) whose goal is to helpbuild reasoning by leaning on a basketball gamerationale. In the BEACONING project presentedby Cardoso, Morgado and Coelho (p. 17), gamifi-cation is used to make lesson plans more fluid.The LudiMoodle project, described by Lavoué (p.19), similarly evaluates the impact of gamificationon motivation.

Motivation and engagement can also be encour-aged by physically involving the learner in thetask. Tangible, hands-on activities can help buildknowledge, and enhance collaborative work andsocial interaction. Hence, compared with purelydigital approaches, hybrid approaches that mix

Introduction to the Special Theme

Educational Technology

by Vassilis Katsouros (Athena Research and Innovation Centre) and Martin Hachet (Inria)


physical and digital components have great poten-tial. The CHILI Lab at EPFL explores approachesbased on robots to support handwriting (Ozgur etal., p. 20) or to accompany a collaborativeproblem-solving task (Nasir et al. p.22). Problem-solving is also the topic of the article by Anastasiouand Ras (p. 23) where tangible user interfaces arecombined with 2D and 3D gestures. InKniwwelino (Maquil and Moll, p. 25), the hands-on activities are supported by a combination of avisual programming language and a physicalmicrocontroller equipped with sensors and dis-plays.

In the EPICSAVE project (Luiz, Lerner andSchnier, p. 26), an immersive room-scaled multi-user 3D virtual simulation environment for med-ical training scenarios enabling realistic and sus-tainable training experiences is presented. TheWEKIT project (Klamma, Koren and Jarke, p. 27)has developed an industrial training platform com-bining sensor technologies with mixed and aug-mented reality for on-the-job training. Virtual, aug-mented and mixed reality is also used in the projectdescribed by Lalos et al. (p. 29) to provide an edu-cational platform for virtual scientific laboratoriesfor STEM education. Kaliakatsos-Papakostas,Kritsis and Katsouros (p. 30) introduce iMusciCA,a web-based workbench that integrates advancedcore enabling technologies, including 3D designand printing of musical instruments, body trackingsensors for gesture recognition, interactive pensand tablets as well as sound generation and pro-cessing tools for STEA(rts)M education. TheWhoLoDancE project (El Raheb and Ioannidis p.32) has developed web-based tools for theanalysis, segmentation, annotation and blending ofdance movements as well as interactive experi-ences that integrate augmented and mixed reality,sonification of movement and visualisation of thehuman body and movement in different avatarsand environments. Leonidis et al. (p. 33), present astudent-oriented and educator-friendly intelligentclassroom that integrates several features, such assynchronous or event-based communication, iden-tification of learners’ behaviour, etc. and attractive,situational environments for learning.

In addition to these research activities and techno-logical developments, various initiatives aim atorganising and teaching EdTech. For example, theEdubase Online Platform (Szilágyi and Berezvai,

p. 35) is a learning-management system (LMS)that provides numerous teaching and testing inter-faces. Romero, Lefèvre and Viéville (p. 36) presentthe SmartEdTech Master program. It is dedicatedto the teaching of, and with, educational tech-nology. Similarly, at a doctoral level, the DE-TELprogram (Fominykh and Prasolova-Førland, p. 38)is a European initiative dedicated to a new form oftechnology-enhanced learning. For all theseresearch works and initiatives in EdTech, ethicsshould remain a central pillar of the upcoming gen-eration of teaching tools and approaches. This isthe focus of a manifesto presented by Laurent,Vaufreydaz and Dessus (p. 39).

The articles in this special theme give a broadoverview of the current state of research and appli-cations and insight into ongoing projects in theeducational technologies field.

Digital enhanced learning is also addressed by therecent DEL4ALL (Digital enhanced learning forAll) project [L1], funded by the European Union.With a forward-looking perspective, it analysesbest-practice, success stories as well as challengesand opportunities offered by the increasing adop-tion of technologies, such as blockchain, artificialintelligence and others. DEL4All is expected toprovide a basis for future research directions andpolicy recommendations to enable the transitionfrom Horizon 2020 to Horizon Europe fundedresearch and innovation projects in the area of dig-ital enhanced learning.

Link:

[L1] https://www.del4all.eu/

Please contact:

Vassilis Katsouros Athena Research and Innovation Centre, [email protected]

Martin Hachet Inria, [email protected]

These systems are results of a combina-tion of multiple disciplines, such ascomputer science, cognitive psychology,human-robot interaction and educationalresearch. Some advantages of ITS arethat they are location-independent,easily accessible, and offer great flexi-bility, allowing students to learn at theirown pace and not to have to rely on rigidclassroom schedules.

Researchers have found that a learningsession can be improved if the teacher isempathetic to the emotions of thelearner. Such improvements could takethe form of an ITS offering help when itdetects that the learner is confused, or byoffering the learner motivating remarkswhen it detects boredom. ITS systemsthat make use of emotions are known asaffective tutoring systems (ATS). ATScombine tutoring strategies and emotionsensing techniques into a single system.Also, evaluations have shown that theycan contribute positively to the user’slearning experience [1].

A disadvantage of these systems is thatthey are designed for teaching a specificsubject to specific users, makingreusability difficult or even impossible.Moreover, the more diverse and compli-cated a course is, the more difficult it isto manage it. In other words, there is aneed for a universal, well-defined struc-ture in order to facilitate the design ofITS. Fortunately, similar problems indifferent domains have been dealt withsuccessfully using a knowledge repre-sentation and reasoning approach. Suchapproaches use declarative logic andlogical formalisms known as action lan-guages in order to formalise the problemand use AI techniques, such as projec-tion and planning, in order to solve it.ITS that make use of this approachalready exist and are known as cognitivetutoring systems. However, these sys-tems are always modelled around a spe-cific course or problem, they don’t take

any preferences of the user intoaccount and none of them have beenused for affective input. In order totackle the user’s preference problem,researchers have created the field of“adaptive learning systems”. Thesesystems use machine learning tech-niques in order to adapt the content of agiven course according to the prefer-ences of their current user. However,these systems do not offer feedback tothe user (emotional or cognitive) as

they are mainly concerned about thecourse’s presentation.

In order to address these limitations ofboth fields, we have built AFFLOG, anadaptive cognitive affective tutoringsystem that uses answer set program-ming and the event calculus action lan-guage [2] in order to represent the var-ious components and actions of an ATS,and perform reasoning tasks such asplanning for creating a course suitable



Α Logic-Based Affective Tutoring System

by Achilles Dougalis and Dimitris Plexousakis (ICS-FORTH)

In science fiction, artificial agents are portrayed as being capable of interacting with and helping

humans. This aid could take the form of holding intelligent conversations and even acting as

teachers and coaches. Some progress has been made in this direction in real life. Indeed, systems

utilising intelligent agents, such as duolingo™, have proven capable of acting as personal tutors.

These “intelligent tutoring systems” (ITS) emulate a human tutor by using AI techniques to adapt

instructions and teaching according to each individual learner’s background and progress but also

guide the learner through an exercise by providing hints and feedback.

Figure�1:�A�course�as�a��directed�graph.�Each�subchapter�is�depicted�as�a�node,�with�one�or

more�tutorials�represented�as�vectors�linking�it�with�other�nodes.�A�chapter�can�be�traversed

either�by�one�tutorial�(e.g.�tutorial8�for�chapter1)�or�by�a�selection�of�smaller�tutorials.�

Figure�2:�Emotional�Strategies�according�to�the�user’s�current�emotion.

to the current user, as well as offeringthe tools to create a new course fromscratch. The main contributions of thiswork are:• Course representation: We represent

a course and its chapters in such away that it can be effectively used byour system while also providing theauthor of the course with flexibility interms of which modalities to use andalso the design, structure and size ofthe course.

• User representation: Which parts of aspecific course the user has learned,his/her preferred learning styles, aswell as his/her emotional stateregarding a course or a part of it.

• The integration of emotions and theuser’s learning style in the tutor’sdecision making process (Figure 2).

We use the term “tutorial” to describethe main building block for a course. Atutorial can be plain text, a picture, anaudio file or a video file. Importantsemantic properties of a tutorial writtenin ASP include the subchapters thatdenote its beginning and end, as well asits modality, relative difficulty, its dura-tion, and how suitable it is for the cur-

rent user according to his/her learningstyle. A “test” is the method the systemuses in order to understand whether theuser understood a part of the course.

A “course” is the material that the tutoruses to teach in one or more “sessions”.It is essentially a collection of tutorialsand tests that are presented to the user.Each course consists of a number of“chapters” and each chapter consists of anumber of “subchapters”. Subchaptersare traversed using tutorials and can bedescribed as nodes in a graph where thetutorials act as the vectors (Figure 2).Once the tutorials that comprise achapter have been presented to the user,the tutor presents him/her with a test todetermine whether the user understoodthe chapter. If the user passes the test, thetutor assumes that the user understoodthe chapter and proceeds to the nextchapter. If not, the tutor assumes that theuser has not understood one or more ofthe tutorials of the particular chapter, butsince it cannot differentiate betweenwhich tutorial was misunderstood andwhich not, all the tutorials of that chapterare labelled as wrong. Consequently, thecourse must be modified using “plan-

ning” by replacing the wrong tutorialswith unused tutorials for that chapter.When all the chapters are completed andall the tutorials are correct, the course hasbeen completed successfully.

The work here represents a part of aPhD in emotional adaptive tutoring sys-tems. Future work includes systemevaluations, linking the system with thesemantic web via the RDF language,and using machine learning methods tooffer more personalised tutoring.

References:

[1] B. Kort, R. Reilly, R. W. Picard:“An affective model of interplaybetween emotions and learning:Reengineering educationalpedagogy-building a learningcompanion, 2001.”

[2] M. Sergot, R. Kowalski: “A logic-based calculus of events”, NewGeneration Computing 4.1 (1985):67-95.

Please contact:

Achilles DougalisFORTH-ICS, [email protected]


Programming tutoring systemProTuS [L1] was developed in theLearner-Computer Interaction lab [L2]at the Norwegian University of Scienceand Technology (NTNU), in coopera-tion with the University of South-Eastern Norway (USN). The system rep-resents an effort to develop an adaptiveand interactive environment that pro-vides personalised learning to students(Figure 1) [1]. ProTuS covers a complexinterplay of learning resources, tasks,instructions, social dynamics, interac-tions, and learning analytics aimed athelping students to learn introductoryprogramming [2]. The system has beenutilised at several universities for over adecade [3].

The current version of ProTuS includesinteractive learning resources from var-ious content providers, developed atseveral universities from Norway, USAand Canada. In addition, the systemoffers adaptive features based on con-tent recommendation, automatic assess-ment and grading.

Interactive learning contentThe interactive learning content (i.e.,resources) included in the system com-prises four types of activity that supportindividual work. Students can practiceand learn programming through the fol-lowing learning resources [2]:Explanations. ProTuS contains readingcontent (i.e., tutorials) on 15 topics that

are aligned with the curriculum pre-sented in the introduction to Javacourse. The objective behind thereading content is to allow students toleverage on their background knowl-edge on procedural programming (nor-mally taught via Python) and progresswith object-oriented programming(OOP). To do so, the content provides acomparison of the syntax and the basicconcepts in Python (procedural) andJava (OOP).

Interactive examples. For each of theproposed topics, the students have thechance to explore different examples,called Program Construction EXamples(PCEX). Each PCEX content starts with

Learning Introductory Programming

with Smart Learning Environment

by Boban Vesin (University of South-Eastern Norway), Katerina Mangaroska and Michail Giannakos (NTNU)

Programming Tutoring System (ProTuS) is an adaptive learning system developed to support introductory

programming. ProTuS utilises a cross-platform architecture that aggregates and harmonises learner

analytics coming from different systems and quantifies learners’ performance through a set of indicators.

ProTuS has been successfully used within universities to support teaching and learning.

a worked-out example with explanationof how to write a code for a particularproblem. The explanations (that arehidden until a student clicks on the linesof interest) are available for almost alllines in the example and they focus onwhy students need to write a code in acertain way or why certain program-ming constructs are used in the code.

Interactive challenges. To allow aseamless connection between theoryand practice, as well as give students thepossibility to try out the given exam-ples, we present a challenge after eachexample. This allows students to solveone or more challenges related to theexample that they previously viewed.

Coding exercises. This is a type of smartcontent that requires students to writecode or complete a given code toachieve a certain goal. Each codingexercise has a problem description anda baseline code. When students submittheir code, it is tested against a set ofpredefined unit tests and the studentreceives automated feedback onwhether the tests were passed or not.

Interactive examples and challenges(Mastery Grids, PSEX) were developedat the University of Pittsburgh, andcoding exercises (PCRS) at theUniversity of Toronto [2].

Scaling up research efforts tosupport introductory programmingAs a cross-platform architecture,ProTuS aggregates and harmoniseslearner-generated data from several sys-

tems into a set of meaningful indicators.Thus, it acts as a learning eco-systemthat leverages on various learning ana-lytics to enhance personalised and adap-tive learning. To address barriers hin-dering the usefulness and efficiency ofan adaptive learning system, we exam-ined ProTuS usability and evaluated itscross-platform learning analytics capac-ities to support personalised and adap-tive learning [1]. The system has beentested in different universitiesthroughout Norway and other Europeancountries, with the results indicatingthat it shows promise for supportingintroductory programming.

In the future, the authors plan to furtherdevelop the learning analytics compo-nent of ProTuS and provide morelearner-centred visualisations andcross-analytics (e.g., data coming fromassignments, project-work, coursegrading, and third-party software) [2].Furthermore, the overall goal is toinvestigate how analytics derived fromvarious sources can help us to constructefficient teaching strategies and supportonline and blended teaching andlearning of introductory programming.

This project is a product of a researchcollaboration between NTNU and USN,receiving funding from NorwegianAgency for International Cooperationand Quality Enhancement in HigherEducation (DIG-P44/2019), theNorwegian Research Council(255129/H20) and NTNU’s ExcellentEducation program (NTNUToppundervisning).

Links:

[L1] https://protus.idi.ntnu.no/[L2] https://lci.idi.ntnu.no/

References:

[1] B. Vesin, K. Mangaroska, M.Giannakos: “Learning in smartenvironments: user-centered designand analytics of an adaptivelearning system,” Smart Learn.Environ.

[2] K. Mangaroska, B. Vesin, and M.Giannakos, “Cross-platformanalytics: A step towardspersonalization and adaptation ineducation,” LAK

[3] A. Klašnja-Milićević, B. Vesin, andM. Ivanović, “Social taggingstrategy for enhancing e-learningexperience,” C&E

Please contact:

Boban Vesin University of South-Eastern Norway,[email protected]



Figure�1:

Personalised�learning

in�ProTuS.


A wide range of technological compo-nents support or facilitate many suc-cessful approaches in the field of educa-tion, such as blended learning or flippedclassroom learning. Learning analyticsand educational recommender systemsare based on statistical models and artifi-cial intelligence; gamification andsocial- and peer-learning require appro-priate backend services; learning con-tent and media of these systems are typi-cally hosted on different servers; andvirtual and augmented reality used ineducational systems require specialisedhardware or software. Many promisingapproaches are being developed as iso-lated solutions, which individually arequite successful, but would only reachtheir full potential when used jointly.Together with the Fraunhofer Academy,Fraunhofer FOKUS is coordinating aresearch project that seeks to integrateisolated solutions with each otherthrough a middleware component.

The Common Learning Middleware[L1] is based on open standards andspecifications for educational technolo-gies, including standardised interfacedefinitions [L2], metadata specifications

for content structures [L3], learningobjects [L4], and quizzes [L5], andstandards for persisting activity data[L6]. Figure 1 shows the overall archi-tecture of the Common LearningMiddleware. In the underlying concep-tual design, every element that can beintegrated into the learning context, beit a text, image, video, dashboard or vir-tual reality, is abstracted as a tool. Thedifferent servers act as tool providersand can publish and subscribe theiroffers to the middleware. From there,the user interfaces, such as traditionallearning management systems (e.g.,Moodle or ILIAS) or advanced learningapplications, can access the toolsthrough standardised interfaces. Themiddleware verifies the roles and rightsof the requesting users via its own userenrolment system for each access. Inaddition, existing user management sys-tems can be connected to the middle-ware to enable cross-system logins.This enables the most diverse learningscenarios, which can be very well tai-lored to the respective learning context,without requiring programming skillsof the content creator. To showcase oursystem, we give a few example sce-

narios, which were realised on the basisof individual solutions.

Together with the institutes FraunhoferFIT and Fraunhofer IML, we have cre-ated a learning offer for the field of datascience in which the learning contentfrom the databases of several ILIASplatforms is presented in a self-devel-oped learning portal from an externalservice provider. The portal also offersprogramming exercises (as individualJupyter notebooks) after each majorlearning unit. Course participants cancommunicate via a tool provided by astart-up´s innovative social learningplatform. In another case, together withFraunhofer IOSB and Fraunhofer FIT,we combined various tools for the cybersecurity domain. It merges onlinelearning content from ILIAS and OpenedX learning platforms with interactivelearning applications developed by anexternal service provider into a singleoffering. In addition, the learned theo-ries can be practically applied in theserious virtual reality game Lost Earth2307 [1], in which the learner has tosolve various security-relevant missionsin a future scenario. The game is seam-lessly loaded into the platform. Themiddleware is also utilised for univer-sity courses in computer science [2].And in a parallel project, we even usedthese interfaces to link six different edu-cational institutes, which normally havelittle contact, from the fields of crafts,computer science and general schooleducation. This created synergiesbetween the institutes on the basis ofcontent and technology. On the onehand, courses on bookkeeping andaccounting only had to be developedonce and could be utilised by all institu-tions in their respective learning plat-forms. On the other hand, moreadvanced components, such as learning

Interoperable Education Infrastructures:

A Middleware that Brings Together Adaptive,

Social and virtual Learning Technologies

by Christopher Krauss and Manfred Hauswirth (Fraunhofer FOKUS)

What should a course provider do if all course content, which is stored in Moodle, needs to be migrated

to a new learning management system? How could a provider easily use advanced technologies like

learning analytics, learning recommender systems or virtual learning to create a compelling learning

experience? How can a provider incorporate the content of another provider into an existing course? To

address such questions, we developed the Common Learning Middleware in a joint project with several

Fraunhofer institutes trying to solve these typical challenges facing educational institutions.

Figure�1:�Simplified�architecture�of�a�learning�infrastructure�based�on�the�Common�Learning

Middleware.

analytics, gamification, learning pathsand a learning recommender system,were realised as tools and offered forthe respective courses via the middle-ware.

Our recommender system [3] is a spe-cial tool provider that plays a centralrole, which focuses on determining“learning needs”. These are numericalvalues that represent the relevance ofthe corresponding content to each indi-vidual course participant. The higherthe individual’s learning need for atopic, the more important it is that thelearner should work on it. The learningplatform loads the recommendationtool, which presents the content recom-mendations for the most relevant topicsin order to make the learning processmore efficient and effective. The rec-ommender takes into account informa-tion that concerns the learning contentitself. This includes, for example, dataon whether certain content is relevantfor exams or whether a lecture is aboutto take place. At the same time, infor-mation collected through direct userinteraction is also included. Forexample, whether the participant hasalready edited all the underlying con-tent, how the participant self-assesses inthe subject area, how he or she has per-formed in exercises or how much of thecontent has already been viewed withthe platform and for how long. A dis-tinctive feature of the recommender

system is the “forgetting effect”, whichprovides information on whether thecontent was supposedly forgotten againbased on the type of media, its scopeand complexity as well as the timeelapsed since the last learning. Theusers of the platform then see generaland thematic recommendations in cate-gories such as: “With these topics youcan prepare for the next lecture.”, “Youhaven't done so well in these exercisesyet.” or “You might have forgotten thisalready.”

The Common Learning Middlewaremakes it possible to combine the mostdiverse educational technologieswithout having to forego content pro-tection and rights management. Contentfrom different learning managementsystems, such as Moodle, ILIAS oropen edX, can easily be merged via acommon interface definition andenriched with innovative technologies,such as learning analytics, gamification,social learning, virtual reality orlearning recommender systems. Variousinstitutions have already tested thismiddleware and Fraunhofer is success-fully using it in several of its courses,such as those of the FraunhoferFOKUS-Akademie [L7]. In response tohigh demand, the middleware is beingopened up to external companies thatvalue the independence of certain solu-tions.

Links:

[L1] https://kwz.me/hKr[L2] https://kwz.me/hKY[L3] https://www.imsglobal.org/cc/[L4] https://www.imsglobal.org/metadata/[L5] https://www.imsglobal.org/question/[L6] https://www.adlnet.gov/projects/xapi/[L7] https://akademie.fokus.fraunhofer.de/

References:

[1] A. Streicher, J. D. Smeddinck:“Personalized and adaptive seriousgames”, Entertainment Computingand Serious Games. Springer,2016. 332-377.

[2] C. Krauss, et al.: “TeachingAdvanced Web Technologies witha Mobile Learning CompanionApplication”, in: Proc. of mLearn2017, ACM, , Larnaca, Cyprus.

[3] C. Krauss, A. Merceron, S.Arbanowski: “Smart LearningObject Recommendations based onTime-Dependent Learning NeedModels”, in Proc. of EDM 2019,M. Desmarais, et al. (eds.), pp. 599- 602, July 2-5, 2019, Montréal,Canada.

Please contact:

Christopher KraussFraunhofer FOKUS, [email protected]

Manfred HauswirthFraunhofer FOKUS, [email protected]



At Fraunhofer Institute for IntelligentAnalysis and Information Systems IAIS,the business unit “Smart Coding andLearning” is dedicated to the topic ofconveying digital skills, such as coding,in a sustainable way, regardless ofgender, age or prior knowledge. As partof this, the educational program“Roberta® – Learning with robots” [L1]has been successfully accompanyingteachers as well as inspiring students inSTEM subjects since 2002. To date

“Roberta” has reached over 450,000children in Germany. This is one ofEurope’s largest STEM initiatives, itsexpansion being due in part to the EUproject “Roberta goes EU” [L4]. Inaddition to the success of its hands-ondidactic concept, one key component ofthe project is the development andimplementation of new technology: theopen-source programming environment“Open Roberta Lab” and the visual pro-gramming language NEPO® in partic-

ular, both initiated with the support ofGoogle.org.

“Open Roberta Lab” is an integratedprogramming environment (IDE) thatis available online, free of charge [L2].As a state-of-the-art and open sourcecloud programming technology (CPT),it enables the web-based programmingof hardware systems, such as robotsand microboards, with the program-ming language NEPO on any computa-

Fraunhofer IAIS IoT Programming Language

NEPO® in the Open Roberta® Lab

by Thorsten Leimbach (Fraunhofer IAIS), Daria Tomala (Fraunhofer IAIS)

Technology now pervades all areas of our lives, including our home life, education and work. As

society becomes increasingly digitalised, digital skills such as “computational thinking” are becoming

more important – this applies to children in school, adults in the workforce and senior citizens alike.


tional device (e.g., smartphone, tablet,PC) and on standard operating systems(e.g., Mac OS, Linux, Windows) inwhich conventional browsers run. Thisindependence of computational devicesand operating systems as well as theweb-based approach itself (i.e., noinstallation is needed), minimises thetechnical obstacles faced by manyusers, especially in educational andworkplace environments, and supportsthe adoption and potential purview ofthe technology.

The enthusiastic reception to OpenRoberta is reflected in the continuouslygrowing number of users worldwidesince its launch in 2014. In 2015 theOpen Roberta Lab recorded 11,000 visi-tors, while in 2017 the website was vis-ited more than 107,000 times. ByOctober 2019 this number more thantripled to a total of 336,270 users from +100 countries.

One key to the success of the platform isNEPO, the intuitive graphical program-

ming language made by FraunhoferIAIS. NEPO (New Easy ProgrammingOnline) reduces the complexity of text-based programming languages, withoutbeing inferior in scope and perform-ance. Despite its block-based approach,NEPO is a full programming languagethat allows the creation of programscontaining any regular coding compo-nents, such as commands, control struc-tures, lists, variables and functions.

The programming principle is easilyunderstood: The coding components inthe form of NEPO blocks are adjustedvia “drag and drop” to a pre-definedstarting point of the program and arethemselves colour-coded to allow therecognition of semantics beforehandand therefore to minimise errors. Theblock-based structure enables syntacticerrors to be avoided altogether.Furthermore, the textual designation ofthe respective blocks makes it possibleto quickly understand the functionalityof the program created. The CPT frame-work also provides online help to indi-

vidual NEPO coding blocks as well asthe ability to integrate short tutorials.

NEPO allows anyone without previousknowledge to code even sophisticatedrobotic systems with various sensorsand actuators in less than five minutes.So far, it has been successfully used bya wide audience: from primary schools,high-schools and universities, to codinghubs and computer clubs for the elderlyand even in industrial and workplacetraining.

As mentioned above, an increasingvariety of robotic systems is program-mable with Open Roberta. What startedas a platform to let users code one robotalone, has now advanced into a multi-system programming solution that letsusers code a wide range of robotic sys-tems intuitively. Starting with educa-tional robots and microcontrollers forchildren aged 7+, humanoid robots,physical computing platforms, smarthome devices or other Internet ofThings (IoT) applications for the per-sonal use, through to potential applica-tions in the industrial sector.

Since NEPO blocks are speciallydesigned by Fraunhofer IAIS to fit anychosen target system and can thereforebe adapted into a customised solution,the areas of application are almost end-less. On the website, the code genera-tion – from NEPO Blocks to textualprogramming languages, such asPython and Java, C – runs completelyon Fraunhofer servers in Germany.However, a copy of the open sourceserver itself can be ported to other sys-tems, e.g., Raspberry Pi [L3].

Figure�1:�the�Open�Roberta�Lab.�On�the�left�the�NEPO�blocks,�on�the�right�the�corresponding�generated�source�code.

Figure�2:�NEPO�blocks�can�be�individually�adapted�to�the�respective�hardware�system.



The functionalities of the Open RobertaLab also include: availability in 17 lan-guages, a gallery to share programsonline, a virtual simulation and theability to describe the programs as wellas to look into the textual source codebehind the NEPO blocks. This combi-nation of features results in a low-threshold and sustainable introductionto coding.

Open Roberta follows the well-provenconcept of “Roberta”, to make abstractconstructs in coding tangible and easilyunderstood by a wide audience. In com-bination with teacher trainings, teachingmaterials and a nationwide network,Open Roberta is becoming increasingly

integrated into classrooms – in the fullspectrum from primary schools to uni-versities. The project, based atFraunhofer IAIS, is now aiming toextend its approach to one of the majorresearch and impact fields of FraunhoferIAIS, artificial intelligence (AI).

Links:

[L1] https://www.roberta-home.de/ [L2] https://lab.open-roberta.org/ [L3] https://kwz.me/hKB[L4] https://kwz.me/hKb

References:

[1] B. Jost, et al.: “GraphicalProgramming Environments forEducational Robots: Open Roberta -

Yet another One?", IEEE MTELworkshop in conjunction with ISM2014.

[2] M. Ketterl, et al: “Open Roberta - aWeb Based Approach VisuallyProgram Real Educational Robots”,LOM, ISSN: 1903-248X.

[3] A. Bredenfeld, T. Leimbach, “TheRoberta® Initiative”, in: Proc. ofSIMPAR 2010, Darmstadt(Workshop Teaching robotics,teaching with robotics), 2010, S. 1-2.

Please contact:

Daria TomalaFraunhofer IAIS, [email protected]

Forgetting is a ubiquitous phenomenonof human cognition that not only pre-vents students from remembering whatthey have learnt before but also hindersfuture learning, as the latter often buildson prior knowledge. Fortunately, cogni-tive scientists have uncovered simple yetrobust learning strategies that helpimprove long-term memory retention:for instance, spaced repetition. Spacingone’s learning means to temporally dis-tribute learning episodes instead oflearning in a single “massed” session,i.e., cramming. Furthermore, carefullyselecting when to schedule the subse-quent reviews of a given piece of knowl-edge has a significant impact on itsfuture recall probability [1].

On the other hand, concerns about the“one-size-fits-all” human learning para-digm have given rise to the developmentof adaptive learning technologies thattailor instruction to suit the learner’sneeds. In particular, recent researcheffort has been put into developing adap-tive and personalised spacing schedulersfor improving long-term memory reten-tion of simple pieces of knowledge.

These tools sequentially decide whichitem (or question, exercise) to ask thestudent at any time based on her paststudy and performance history. Byfocusing on weaker items, they showsubstantial improvement of the learners’retention at immediate and delayed testscompared to fixed review schedules.Some popular flashcard learning tools,such as Anki and Mnemosyne, use thistype of algorithm.

However, adaptive spacing schedulershave historically focused on pure mem-orisation of single items, such as foreignlanguage words or historical facts. Yet,in real-world educational settings, stu-dents also need to acquire and apply aset of skills for a long period (e.g., inmathematics). In this case, single itemspotentially involve several skills at thesame time and are practiced by the stu-dents to master these underlying skills.With this ongoing research project, weaim at developing skill review sched-ulers that will recommend the right itemat the right time to maximise the proba-bility that the student will correctlyapply these skills in future items. An

example skill reviewing policy from anadaptive spacing algorithm is given inFigure 1.

To address this issue, we chose tofollow a model-based approach: a stu-dent learning and forgetting modelshould help us to accurately infer theimpact on memory decay of selectingany skill or combination of skills at agiven timestamp. Previous work fromLindsey et al. [2] had similarly chosento select the item whose recall proba-bility was closest to a given threshold:in other words, recommending the itemthat is on the verge of being forgotten.Unfortunately, no student model fromthe scientific literature was able to inferskill mastery state and dynamics whenitems involve multiple skills.

To bridge this gap, we developed a newstudent learning and forgetting statis-tical model called DAS3H [3]. DAS3Hbuilds on the DASH model fromLindsey et al. [2]. DASH representspast student practice history (itemattempts and binary correctness out-comes) within a set of time windows to

Modelling Student Learning and Forgetting

for Optimally Scheduling Skill Review

by Benoît Choffin (LRI), Fabrice Popineau (LRI) and Yolaine Bourda (LRI)

Current adaptive and personalised spacing algorithms can help improve students’ long-term memory

retention for simple pieces of knowledge, such as vocabulary in a foreign language. In real-world educational

settings, however, students often need to apply a set of underlying and abstract skills for a long period. At the

French Laboratoire de Recherche en Informatique (LRI), we developed a new student learning and forgetting

statistical model to build an adaptive and personalised skill practice scheduler for human learners.


predict future student performance andis inspired by two major cognitivemodels of human memory, ACT-R andMCM. Most importantly, DASH hasbeen used in a real-world experiment[2] to optimise reviewing schedules forSpanish vocabulary learning in anAmerican middle school. Compared toa fixed review scheduler, the person-alised DASH-based algorithm achieveda 10% improvement at a cumulativeexam at the end of semester. Our modelDAS3H extends DASH by incorpo-rating item-skills relationships infor-mation inside the model structure. Thegoal was both to improve model per-formance when prior information is athand and to use the model estimates ofskill mastery state and memorydynamics it encapsulates to prioritiseitem recommendation.

To evaluate our model’s potential use inan adaptive and personalised spacingalgorithm for reviewing skills, it wasnecessary to first compare its predictivepower to that of other student models onreal-world data. Thus, we testedDAS3H on three large educationaldatasets against four state-of-the-artstudent models from the educationaldata mining literature. We split the stu-dent population into five disjointgroups, trained all models each time onfour groups and predicted unknown stu-dent binary outcomes on the fifth group.On every dataset, DAS3H showed sub-stantially higher predictive performancethan its counterparts, suggesting thatincorporating prior information aboutitem-skills relationships and the forget-ting effect improves over models thatconsider one or the other. Our Python

code has been made public on GitHub:see [L1].

Besides its performance, our modelDAS3H has the advantage of beingsuited to the adaptive skill practice andreview scheduling problem that we aretrying to address. For instance, it couldbe used at any timestamp to predict theimpact of making the student revieweach skill and then to select the mostpromising skill or combination of skills.In future research, we intend to investi-gate heuristics and data-driven policiesfor selecting items to maximise long-term memory retention on a set ofunderlying skills.

Link:

[L1] https://kwz.me/hKt

References:

[1] N. J. Cepeda et al.: “Spacing effectsin learning: A temporal ridgeline ofoptimal retention”, Psychologicalscience, 19(11), 2008

[2] R. V. Lindsey et al.: “Improvingstudents’ long-term knowledgeretention through personalizedreview”, Psychological science,25(3), 2014

[3] B. Choffin et al.: “DAS3H:Modeling Student Learning andForgetting for Optimally SchedulingDistributed Practice of Skills”,Educational Data Mining, 2019

Please contact:

Benoît ChoffinLaboratoire de Recherche enInformatique (LRI), [email protected]

Time t

ALICE

Skill 1(1st acquisition)

Skill 2(1st acquisition)

HeuristicsData-driven policiesOther policies

sictisruHetaDa idr

sili

lliSk 1(1st itisuqac

n)ioit

Tim

taDa - vidrrheOt lipo

t

ne seicliposeicli

A CLI

EC Sk(1st qac

lliSk 2n)ioitisuq

eTim e t

Figure�1:�Example�skill�reviewing�schedule�from�an�adaptive�spacing�algorithm.�Solid�line

indicates�the�first�acquisition�of�a�skill�and�the�dashed�line,�the�subsequent�reviews�of�that�skill.

Alice�is�trying�to�master�two�skills�(resp.�in�orange�and�green)�with�an�adaptive�spacing�tool.

She�first�reviews�skill�1�but�either�because�she�already�forgot�it�or�because�she�did�not�master�it

in�the�first�place,�she�fails�to�recall�it�and�is�rapidly�recommended�to�review�it�again.�The

situation�is�different�for�skill�2:�her�first�attempt�is�a�win�and�the�algorithm�makes�her�wait

longer�before�reviewing�it�again.�Diverse�methods�can�be�used�to�personalize�spacing:

heuristics,�data-driven�reviewing�policies,�etc.

The goal of SIDES 3.0 is to empowermedical students and teachers in dataanalytics to enable them to take chargeof their own progress monitoring andtraining.

We have followed the Semantic Web [1]and Linked Data [2] standards for thesemi-automatic and modular construc-tion of OntoSIDES [3], a knowledgegraph comprising a lightweight domain

ontology serving as a pivot high-levelvocabulary of the query interface withusers, and of a huge dataset that is auto-matically extracted from SIDES dumpsusing mappings. Both the ontological

Ontology-based Learning Analytics in Medicine

by Fabrice Jouanot (Univ. Grenoble Alpes), Olivier Palombi (Univ. Grenoble Alpes, CHU de Grenoble) andMarie-Christine Rousset (Univ. Grenoble Alpes, IUF)

Through SIDES 3.0, we are developing an ontology-based e-learning platform in medicine to make

learning analytics transparent and explainable to end-users (learners and teachers). In this project, the

educational content, the traces of students' activities and the correction of exams are linked and

related to items of an official reference program in a unified data model. As a result, an integrated

access to useful information is provided for student progress monitoring and equipped with a powerful

query language allowing users to express their specific needs relating to data exploration and analysis.

statements and the factual statementsare uniformly described in RDF format[L1], which makes it possible to queryOntoSIDES using the SPARQL querylanguage [L2].

It is important to note that no personaldata concerning students are extractedfrom SIDES. Instances of students inOntoSIDES are just represented byidentifiers of the form sides:etu12402(as shown in Figure 1).

The current version of the OntoSIDESknowledge graph has scaled up to 6,5billion RDF triples describing trainingand assessment activities performed bymore than 145,000 students over almostsix years on the SIDES French nationale-learning platform. In OntoSIDES,exams and training tests are made up ofmultiple choice questions and studentactivities are described at the granularityof time-stamped clicks of students’selected answers to each question.

Through SPARQL queries, students canexplore parts of the program that theyhaven’t yet studied, or launch a compar-ative analysis of their own progress in aspecific part of the program. The samemechanisms let teachers analyse thestrengths and weaknesses of their classcompared to other groups at the samestudy level in other universities.

Since we do not expect end-users tomaster the raw syntax of SPARQL and

to express directly complex queries inSPARQL, we have designed a set ofparametrised queries that users can cus-tomise through a user-friendly inter-face. Figure 1 shows the interface forstudents who can choose to launch oneof the queries in the left column thatwill be parametrised by the student’sidentifier (sides:etud12402 in theexample). The chart in the right side ofthe figure shows the result returned tothe first query for that student, i.e., theaverage results obtained per specialtyby the student (to all the questionshe/she answered related to this spe-cialty), compared to the overall averageobtained per specialty by the wholegroup of active students.

This methodology is not specific tomedicine ; it can be applied to other dis-ciplines for enriching or buildinglearning management systems. Theeffort required depends mainly on theexistence of a reference standard for theeducational program in the target disci-pline.

SIDES 3.0 is an ongoing project fundedby the French Programme Investissementd’Avenir (PIA) through the ANR callDUNE (Développement d'UniversitésNumériques Expérimentales). It is con-ducted under the authority of UNESS(Université Numérique en Santé et enSport) and involves UniversitéGrenoble Alpes, Inria, and EcoleNormale Supérieure as partners. It is

built upon the SIDES e-learning plat-form, which has been used since 2013by all medical schools in France foronline graduation and assessment.

Links:

[L1] http://www.w3.org/TR/rdf-concepts/[L2] https://www.w3.org/TR/

References:

[1] D. Allemang, J. Hendler:“Semantic Web for the WorkingOntologist: Modeling in RDF,RDFS and OWL”, MorganKaufmann, 2011.

[2] T. Heath, C. Bizer: “Linked Data :Evolving the Web into a GlobalDataSpace”, Morgan and Claypool,2011.

[3] O. Palombi, et al.: “OntoSIDES:Ontology-based student progressmonitoring on the nationalevaluation system of FrenchMedical Schools”, ArtificialIntelligence in Medicine, vol 96,2019.

Please contact:

Marie-Christine RoussetUniv. Grenoble Alpes, [email protected]



Figure�1:��Parametrized

queries�for�students.


Universities and other educational insti-tutions confirm degrees and achieve-ments of certain learning outcomes byissuing certificates, which, to date, havelargely been issued on paper or otherphysical formats. Certificates includeseveral statements, the most importantbeing: the qualification or academictitle attained, the name and address ofthe issuer organisation, the name andsignature of the certifier who has vali-dated the facts and is certifying that thequalification is true, the name of thelearner and the date of the examination.Some certificates include additionalstatements, such as the period ofvalidity.

Students usually receive the certificateas a paper document or PDF. This kindof document has the disadvantage ofbeing difficult to verify, and issuinginstitutions need to maintain a registryor database for certificates for a longperiod of time [1]. Employers can onlycheck a certificate’s authenticity byrequesting confirmation from theissuing organisation, which is time-con-suming and expensive, and thus oftenavoided. Consequently, a billion-dollarindustry in fraudulent degrees hasdeveloped [2], with fake educationalcertificates being available for purchaseonline. Certification authorities, univer-sities, and educational institutions haveexpressed a need for a secure and intu-itive solution. Digital standardised doc-uments and blockchain technology

seem ideal to solve many of the above-mentioned problems.

SolutionBlockchain for Education is a web-based platform that supports counterfeitprotection as well as secure manage-ment and easy verification of certifi-cates according to the needs of educa-tional institutions, learners, andemployers. It ensures greater efficiencyand improved security for certificationauthorities through digitisation of cur-rent processes, issuing and registeringof certificates in a blockchain as well asautomatic monitoring of certificates. Tofollow the Industrie 4.0 approach [L1]the platform supports machine-readablecertificates using an extension of OpenBadges [L2].

Our solution supports the wholeprocess from the administration of cer-tificate templates and examination datato the registration of signed certificatesin the blockchain to the final check ofthe validity of the certificate by an HRdepartment, as illustrated in Figure 1. Itshould be noted that the informationregistered in the blockchain does notinclude the learner’s personal data, i.e.the protection of privacy for learners isensured. Furthermore, the platformenables certification authorities torevoke certificates. This could be nec-essary when plagiarism has beendetected, misconduct of the certifiedlearner was proven, or learners have

not provided evidence of requiredrefreshment trainings.

Learners can securely manage their cer-tificates and degrees in their wallet viaan intuitive App running on smart-phones, tablets, and PCs. The walletvisualises the imported certificates,monitors certificates with a time-lim-ited validity and indicates expirations,and provides an easy means to sharecertificates with potential employers.Learners can also print the certificate.The QR-code on the document containsthe verification URL.

Trustworthy verification of certificatesis guaranteed for employers by an easyto use web-based proof service. Whenemployers receive the certificate, theycan upload it by drag and drop on theproof service or scan the QR-code witha smartphone. The verification serviceverifies the given data in theblockchain, performs the proof oforigin, and presents the check result. Inaddition to verifying individual certifi-cates, human resources departments cangain verification overviews ofemployees’ certificates. For example inthe financial sector this simplifies andaccelerates proceedings towards theregulatory authority.

The Blockchain for Education platformis based on the Quorum Blockchain withconsensus algorithm Istanbul BFT,including two smart contracts, the

Blockchain for Education: Lifelong Learning Passport

by Wolfgang Prinz, Sabine Kolvenbach and Rudolf Ruland (Fraunhofer FIT)

Certificates are important to document educational attainments, and educational documentation is required

when it comes to job-seeking and career advancement. It is therefore vital that these records are stored in long-

term available and tamper-proof ledgers. The Blockchain for Education platform is a secure and intuitive solution

for issuing, sharing, and validating of certificates.

Figure�1:�Certification�process�supported�by�Blockchain�for�Education.

InterPlanetary File System (IPFS), adocument management system, and thewallet. The first smart contract registersthe identities of certification authoritiesand certifiers, the second smart contractregisters the fingerprints of the certifi-cates. The IPFS holds profile informa-tion of certification authorities. The doc-ument management system for issuinginstitutions manages the actual payloadof certificates and the parties involved inthe process, namely accreditation andcertification authorities, certifiers,learners, and human resources depart-ments. The API enables the integrationwith existing infrastructures such as thelearning management system Moodle.

Product launchThe Blockchain for Education platformwas developed in an iterative and incre-mental process in strong collaboration

with certification authorities and uni-versities [3]. A first platform prototypewas successfully demonstrated onCEBIT 2018. We are continuouslyextending and enhancing the system. InDecember 2019 we will launch theproduct; our first customer is theFraunhofer Personnel CertificationAuthority [L3]. In parallel we initiatedDigiCerts [L4], a network of universi-ties and educational institutions thataims to collaboratively develop stan-dards and solutions in certificationprocesses. The network is an open net-work and we invite other members fromthe ERCIM community to join it.

Links:

[L1] https://kwz.me/hKd[L2] https://openbadges.org/[L3] https://kwz.me/hKc[L4] https://www.digicerts.eu/

References:

[1] A. Grech, A. F. Camilleri:“Blockchain in Education”, No.JRC108255, Joint Research Centre(Seville site), 2017.

[2] A. Ezell, J. Bear: “Degree mills:The billion-dollar industry that hassold over a million fake diplomas”,Pyr Books, 2005.

[3] W. Gräther, et al.: “Blockchain forEducation: Lifelong LearningPassport”, ERCIM BlockchainWorkshop 2018: BlockchainEngineering: Challenges andOpportunities for Computer ScienceResearch, Amsterdam 2018.doi:10.18420/blockchain2018_07.

Please contact:

Sabine Kolvenbach, Fraunhofer FIT, [email protected]



Formal action languages are well-estab-lished logical theories for reasoningabout the dynamics of changing worlds,contributing solutions to domains asdiverse as high-level robot cognition,ambient intelligence and complex eventdetection. The heterogeneous applica-tion areas of these formalisms attract theinterest of researchers from a variety ofdisciplines and with diverse back-grounds. Given the quite advanced tech-nical training needed to master logic-based reasoning, the educational processof understanding both the theoreticaland practical aspects of common-sensereasoning is a challenging task.

EduBAI [1] (Educational Basketballplaying using Artificial Intelligence)aims to facilitate the learning process oflogic-based reasoning. EduBAI is agame platform that assists students andresearchers having basic knowledge ofthe underlying formalisms in writing,executing and evaluating logic-basedaxiomatizations. It offers a testbed forexperimenting with a rich range of rea-

soning features, structured around agame scenario involving three-playerteams competing against each other ina basketball match. The gameplaygives the ability to explore aspectsrelated to causality-based reasoning,geospatial reasoning, multi-playercoordination, temporal reasoning,probabilistic reasoning, function opti-misation and others.

The platform was developed in the con-text of a post-graduate knowledge rep-resentation and reasoning course at theComputer Science Department of theUniversity of Crete and has been co-developed by students. The gamifica-tion nature of the approach aims tooffer a more entertaining way of intro-ducing logic-based programming,motivated by the observation that toyproblems in the form of grid settingsand logic puzzles have proven to be aneffective way of understanding insightsof AI programming. The platform isfree-to-use and available online fortesting and play [L1].

Users of the platform model the intelli-gence of their team using formal lan-guages, such as the Event Calculus andAnswer Set Programming (ASP), in thecontext of a dynamic, non-deterministicdomain. The Event Calculus is a narra-tive-based many-sorted first-order lan-guage for reasoning about action andchange, while ASP is a declarativeproblem-solving paradigm orientedtowards complex combinatorial searchproblems.

Each of the six players can move withinthe premises of the EduBAI basketballcourt, represented as a 5 × 7 grid(Figure 1). The players can performonly a limited set of actions, namelyshoot, move (up, down right, left) andpass. Player actions are contingent onappropriate preconditions; for instance,a player can perform a pass only if sheis in possession of the ball. Moreover,the outcome of each action is decidedprobabilistically, e.g., a shoot action has95% probability of being successful if itoccurs in the same cell as the basket;

An Educational Platform for Logic-based Reasoning

by Dimitrios Arampatzis (FORTH-ICS), Maria Doulgeraki (FORTH-ICS), Michail Giannoulis (Univ. of Crete), Evropi Stefanidi (FORTH-ICS), and Theodore Patkos (FORTH-ICS)

EduBAI (Educational Basketball playing using Artificial Intelligence) is an educational platform that helps

users familiarise themselves with the main tenets of common-sense reasoning in dynamic, causal domains,

by means of an interactive entertaining environment. This article discusses the design and features of the

platform, along with the rationale of sample game tactics of diverse modelling complexity.


distance from the basket reduces theprior probability by 17% per cell,whereas each opponent in the same cellreduces the probability by 25%. Thegame proceeds in rounds. At the begin-ning of each round, the human usersdecide the starting position of theirplayers, as well as the defending andattacking tactic that their team willfollow during the round; these tacticscannot change in the course of the givenround, therefore once this initialisationstep is completed, the game proceedswithout any human user intervention.

One of the objectives of the EduBAIplatform is to help users understand themain principles of causality-based rea-soning, through the axiomatization oftheir own attacking and defending teamtactic intelligence. EduBAI enables theuser to experiment with features ofincreasing complexity, still presentingan operational team every time, permit-ting a gentle introduction to the model-ling concepts. Tactic encodings can startby ignoring the opponent positions, forinstance, and only follow specific, pre-defined patterns; or they may considerspatial and topological relations amongplayers, probabilistic reasoning on theeffects of actions, cost function optimi-sations etc.; they may even take intoaccount past actions, in order to learnpatterns and predict future moves.

EduBAI consists of multiple web serv-ices that expose their functionalityeither via a Rest API or a web-socketconnection. The core components of theplatform involve: the User and UserStatus Services, nodeJS services that

store, manage, and serve data aboutusers and their game sessions; theTactics Service, that manages customtactics and provides a Restful APIdeveloped using nodeJS that allows theuploading of ASP code; the PlayersPositioning Service for real time con-nection between users, using nodeJSand webSockets; the Game System, aJava web service responsible for run-ning a game session and for calling theClingo ASP solver to execute reasoningtasks; and, the Graphical User Interface,which is Web-based and implementedusing HTML, CSS and JavaScript,designed with respect to human-com-puter interaction (HCI) principles.

Future plans involve extending the plat-form with friendly interactive featuresthat will further enhance the educationalexperience - for example, an embeddededitor to enable users to encode theirtactics or a tutorial-style assistant to

walk a new user through the main func-tionalities. In addition, we plan to offera more advanced competition environ-ment, where the users can schedulematches or upload their own AI-enabledbots and keep track of online statistics.

Link:

[L1] http://139.91.183.97:8181/edubai/

Reference:

[1] D. Arampatzis, et al.: “EduBAI: AnEducational Platform for Logic-Based Reasoning”, in: ArtificialIntelligence Applications andInnovations, AIAI 2019, IFIPAdvances in Information andCommunication Technology, pp.464-472 vol 559, 2019.

Please contact:

Theodore Patkos, DimitriosArampatzis, FORTH-ICS, [email protected], [email protected]

Figure�1:�Snapshot�of�the�EduBAI�UI.

Authoring Game-Based Learning Activities

that are Manageable by Teachers

by Pedro Cardoso (FEUP/ FBAUP and INESC TEC), Leonel Morgado (Universidade Aberta and INESC TEC),and António Coelho (FEUP and INESC TEC)

The great ambition of using games as the cornerstone of education is hindered by its associated teaching

workload. The BEACONING project developed a framework based on an authoring tool for gamified lesson

paths, which has been rolled-out in large scale across Europe. It includes stages for planning game-based

educational activities, plus their deployment, monitoring, and assessment.

Games are great tools for learning and assuch many teachers wish to employthem in their teaching practices.Notwithstanding, using games in educa-tion is hard. It is hard to plan the time

and the teachers’ tasks; the students’activities; to keep track of what each stu-dent is doing; to assess and providefeedback, and so on – and all theseobstacles encumber and limit educa-

tional adoption of games [1]. The BEA-CONING project [L1] acknowledgesthis and created the concept of a gami-fied lesson path [2], by linking gameplots (the narrative and level design)

provided by game development compa-nies with educational activities within atriadic assessment model [3]. Using anauthoring tool developed at INESC TEC(Figure 1), a learning designer creates agamified lesson path by selecting agame plot, linking learning activitiesstructured as missions and quests into it,and finally by associating those activi-ties with specific learning goals andchallenges (mini-games). Learningdesigners are either experts at creatingeducational content, within companiesand organisations, or teachers withtraining in this area. By selecting a gam-ified lesson path, teachers can automati-cally deploy games to their students,adapting them to specific individualrequirements.

Not only students receive the games aspart of educational activities from theirteachers, but this deployment is alsolinked to the learning management plat-forms: teachers can track which studentshave not yet started learning within agamified lesson plan, which ones arecurrently amidst it and where, and whichstudents have finished it and how.Students can also track which teachersand courses have assigned them gami-fied lesson plans, and how they are pro-gressing in each.

All of this linking is done viaanonymization, using the BEACONINGplatform as middleware, so that gamingcompanies can deploy games to indi-vidual students over the Web withoutactually knowing who each student is:individual details are kept within the

learning management system andretrieved directly by the game on stu-dents’ smartphones, without being sentto the companies or even to the middle-ware. BEACONING provides each stu-dent’s game with information on theactual web call to retrieve informationsuch as player names, but this transit ofinformation occurs solely between thestudent’s smartphone and the school’slearning management system.

By using BEACONING to track stu-dents’ progress, teachers can bettermanage their time and effort allocationwhen employing videogames in school,thus reducing their workload andempowering them to use these as a reli-able and regular pedagogic approach“anytime and anywhere”. Thisauthoring and deployment pipelinehave now been successfully rolled outin large-scale testing across Europe(Figure 2) [L2].

BEACONING’s prototype andauthoring pipeline approach holds thepotential for application in many othernon-traditional learning activities.Beyond gaming, we are exploring andexpanding this approach to enablewidespread deployment of activelearning, location-based learning,immersive environments, outdoor activ-ities, and more.

Links:

[L1] https://www.beaconing.eu/[L2] https://beaconing.eu/wp-content/uploads/deliverables/D6.3.pdf

References:

[1] B. B. Marklund, A. S. A. Taylor:“Educational Games in Practice:The Challenges Involved inConducting a Game-BasedCurriculum”, Electronic Journal ofe-Learning, 14(2), 122-135, 2016

[2] A. Bourazeri, et al.: « Taxonomy ofa Gamified Lesson Path for STEMEducation: The BeaconingApproach”, in European Conf. onGames Based Learning (pp. 29-37). Academic ConferencesInternational Ltd, 2017.

[3] R. Baptista, A. Coelho, C. V. deCarvalho: “Relation BetweenGame Genres and Competences forIn-Game Certification”, in Int.Conf. on Serious Games,Interaction, and Simulation (pp.28-35). Springer, 2015.

Please contact:

António Coelho FEUP & INESC TEC, [email protected]

Leonel Morgado Universidade Aberta & INESC TEC,[email protected]



Figure�1:�Gamified�Lesson�Plan�with�branching�activities,�viewed�in�the�authoring�tool�[L2].

Figure�2:�Interface�of�a�location-based

activity�used�in�pilot�testing�of�the

BEACONING�platform.


The gamification field has been growingover the last decade. Most studies thathave investigated the use of gamifica-tion in education have highlighted itsbenefits in terms of learner performanceand motivation. However, these studieshave generally focused on short-termeffects, and do not differentiate betweenthe impact of each game elementaccording to the context or the learnerprofile. More recent studies, consideringthe impact of each game element sepa-rately, have shown that their impactdepends on the learner profile [1] (e.g.,motivation, player type, personalitytraits), with some game mechanismsbeing detrimental to specific learners[2]. For instance, in a previous study, weshowed that amongst the most engagedlearners (i.e., learners who use the envi-ronment the longest), those with adaptedgame elements spend significantly moretime in the learning environment [3].

In this context, the LudiMoodle [L1]project aims to acquire new insights intothe impacts of specific game elements,including: timer; leaderboard; progress;avatar; badges and scores, on learnermotivation when using a digital learningenvironment (Figure 1). The project willprovide (i) researchers with a model for

the adaptation of game elements to thelearner profile; (ii) instructionaldesigners with a generic plugin togamify resources offered in the Moodlelearning environment; and (iii) teacherswith recommendations to gamify theircourses in a way that motivates learnerswith different profiles and motivation.

The project is led by the University ofLyon [L2]. It involves the educationauthority (Rectorat) of Lyon [L3],researchers both in computer sciences(LIRIS lab [L4]) and educational sci-ences (ECP lab [L5]), instructionaldesigners from the University Lyon 3[L6] for the design of digital learningresources in close collaboration withteachers, as well as the Edunao [L7]company dedicated to the developmentof the game elements. Ground experi-ments are conducted in middle schoolsof the Lyon Educational District.Teachers are involved in the design ofgame elements during participatorydesign sessions to create meaningfuland motivating game elements.

The four-year project began in January2017. We ran a first experiment inMarch-April 2019 over ten course ses-sions that involved 258 fourth grade

pupils in four middle-schools, in amathematics course. Learners usedindividual tablets that displayed exer-cises and they were guided in the digitalenvironment. The six game elementswere randomly assigned to learners, onegame element per learner, to answertwo research questions: (i) How doesgamification influence learner motiva-tion? And (ii) Which factors influencethe impact of game elements on learnermotivation? We used questionnaires toassess the initial level of motivation ofthe learners before the course and thelevel of motivation at the end of thecourse. We also asked learners to fulfillthe Hexad [L8] questionnaire to identifytheir player type. Finally, we collectedall interaction traces with the learningenvironment to identify their perform-ance (correct and incorrect answers)and engaged behaviours (e.g., numberof exercises, time spent on an exercise,time spent to answer a question).

The first results show that learners’intrinsic and extrinsic motivationdecreased during the experiment. Theseresults are not surprising since gameelements were randomly distributedamongst learners and may not fit theirmotivation for the course and prefer-ences for game mechanisms. We alsoobserved an increase in learner motiva-tion in the students that were least moti-vated initially. A deeper analysisshowed that two factors influenced theimpact of game elements on learnermotivation and performance: (i) the ini-tial level of motivation at the beginningof the course and (ii) the player profile,achiever and player dimensions beingthe most important. An analysis pergame element also showed that (i) eachgame element influences differentdimensions of motivation (eitherintrinsic, extrinsic or amotivation); and(ii) the impact of each game elementdepends on a different combination offactors including the initial motivationand certain dimensions of the playerprofile.

LudiMoodle: Adaptive Gamification to Improve

Learner Motivation

by Élise Lavoué (Université Jean Moulin Lyon 3)

The LudiMoodle project aims to adapt game elements integrated into the Moodle learning

environment to help motivate learners.

Figure�1:�Game�elements�developed�within�the�LudiMoodle�project�for�the�Moodle�learning

platform.

A second experiment is planned forMarch-April 2020 to identify the impactof adaptive gamification on learnermotivation. An adaptation engine willbe integrated into the digital learningenvironment. Static adaptation ruleswill be defined according to the factorsidentified in the first experiment, to pro-pose game elements adapted to eachlearner’s profile at the beginning of thecourse. We will also define a dynamicadaptation process to suggest game ele-ments according to engaged or disen-gaged behaviours observed during thecourse via learners’ interaction traceswith the learning environment. At theend of the course, we will compare theimpact of adapted game elements onlearner motivation compared to gameelements attributed randomly. We willalso compare the impact of adaptedgame elements depending on the type of

adaptation (static vs. dynamic). Webelieve these studies will provide newinsights into the impact of adaptivegamification on learner motivation, andrecommendations for designers andteachers to adapt game elements tolearners.

This work is a part of the LudiMoodleproject financed by the e-FRANProgramme d’investissement d’avenir,operated by the Caisse des Dépots.

Links:

[L1] https://ludimoodle.universite-lyon.fr[L2] https://kwz.me/hKy[L3] http://www.ac-lyon.fr/[L4] https://liris.cnrs.fr/en/team/sical[L5] https://kwz.me/hKH[L6] https://kwz.me/hKE[L7] https://www.edunao.com/en/home/[L8] https://www.gamified.uk/user-types/

References:

[1] S. Hallifax, et al.: “Factors toConsider for TailoredGamification”, CHI PLAY’19,2019, pp. 559-572.

[2] S. Hallifax, et al.: “Adaptivegamification in education: Aliterature review of current trendsand developments”, EC-TEL 2019,Springer, pp. 294–307.

[3] E. Lavoué, et al.: “AdaptiveGamification for LearningEnvironments”, IEEE TLT, 12, 1,pp. 16-28, 2018.

Please contact:

Élise LavouéUniversité Jean Moulin Lyon 3, [email protected]



Research shows that having richer sen-sory information enhances visual per-ception and visual-motor coordination[1]. To let children experience a range ofsensory information during letterlearning, teachers use techniques such asdrawing letters in sand-filled boxes ortouching the letters carved in a piece ofwood. The same principle is adopted bytherapy centres for the development oflanguage skills, such as the “draw onyour back” game. During the game, achild and a therapist take turns indrawing a letter on the other’s back withtheir finger and guessing what is drawn.

Similarly, robots can be used as tangibletools enhancing sensory information, aswe demonstrated in a previous study thatused the Cellulo robots to help pre-schoolers learn the grapheme and theductus of letters [1]. Cellulo is a palm-sized, haptic-enabled robot developed atEPFL [L1] [2] that can move and be

moved on “maps”, i.e., printed sheets ofpaper covered with a dotted-pattern thatenables accurate localisation (Figure 1shows children interacting with theirrobots on letter maps). Cellulo robotsare versatile, easy to set up and well-suited for classroom activities. They arecontrolled by an application running ona computer or a tablet and feature var-ious interaction modalities; many suchrobots can also be simultaneously usedas a “swarm” in an activity [3].

Our goal, inspired by interactions withteachers and therapists, was to design amodular, highly engaging, highly adapt-able robot-assisted activity to help chil-dren with attention and/or visuo-motorcoordination issues in learning thegrapheme and ductus of cursive letters.

Through several iterations within aschool and a number of therapy centres,we designed an activity composed of

three sub-activities targeting differentaspects of letter handwriting learning.The setup envisions multiple kids sit-ting at a table, each with a Cellulo and amap displaying the grapheme of theletter, as shown in Figure 1. The sub-activities are:1. “Watch the Robot”: Cellulo

autonomously moves along the let-ter’s grapheme on the map, followingthe ductus. The child only watchesthe robot and the letter’s phoneme isplayed at the beginning and end ofthe robot’s writing, to strengthen thelink with the correspondinggrapheme and ductus.

2. “Feel the Robot”: We ask the child toput their hand on Cellulo while itmoves along the letter’s grapheme topassively feel the motion of the robot.Experiments with children withvisio-motor coordination difficultieshighlighted the importance of thisactivity: the child has to adjust the

Can Tangible Robots Support Children

in Learning Handwriting?

by Arzu Guneysu Ozgur, Barbara Bruno, Thibault Asselborn and Pierre Dillenbourg, (EPFL)

A large body of research suggests that robots could indeed be useful for supporting children in learning

handwriting. However, few studies have investigated the role and use of tangible robots in teaching

handwriting to children with attention and/or visuo-motor coordination difficulties. Over the course of

multiple iterations, globally involving 17 typically developed children and 12 children with attention and

visio-motor coordination issues within one school and two different therapy centres, we have designed

a robotic activity to teach the grapheme (shape) and the ductus (the way to draw) of cursive letters.


force applied on top of the robot,since applying excessive force pre-vents the robot from moving.

3. “Drive the Robot”: The child activelydrives Cellulo along the letter’sgrapheme, trying to repeat the ductusseen previously. Although the childhas full control over the robot’smovements, Cellulo provides hapticfeedback, trying to lead the child tothe right path if it is moved awayfrom it. Additionally, Cellulo’s LEDsare blue when the correct path is fol-lowed, and red otherwise. This visual+ haptic feedback helps the childrecognise and avoid errors.

The design includes a final teamactivity in which three to four children,alternatively playing as one “writer”and many “guessers”, sit side-by-side atthe table, with a barrier in between pre-venting the guessers from seeing thewriter. Using the map as a reference, the“writer” is asked to write a letter usingCellulo as done in “Drive the Robot”.The movements of the “writer” robotare replicated in real-time by the“guesser” robots, autonomouslymoving on a blank map placed in frontof each “guesser”. The goal of the gameis for the “guessers” to successfullyguess the letter drawn by the “writer”.During one iteration, therapists sug-gested to adapt this game to be morechallenging for the “writer”: to this end,we tested the case in which the “writer”is also given a blank map and has toremember both the proper graphemeand ductus for the letter that they areinstructed to write. The game thusbecomes a team effort in which wrong

guesses from the “guessers” are fedback to the “writer” and the goal is forthe team to make the “guessers” cor-rectly guess the letter with the leastnumber of attempts. This adaptationalso allows us to track the progress ofthe child in handwriting, by analysingthe “writer” robot’s trajectory.

The map also proved to be a crucialadaptation element: different institu-tions use slightly different graphemesfor the same letter, as well as differentcues for the initial strokes. For thisreason, we designed multiple variantsof the maps (e.g., featuring car-racingelements to introduce a gaming dimen-sion or skate ramps and sea waves torelate to one way in which specificstrokes are taught).

Finally, therapists highlighted theimportance of knowledge transfer, i.e.,ensuring that children who learn theappropriate ductus and grapheme withthe robot are capable of writing equallywell with pen and paper. Thus, in oneiteration we interspersed pen-and-paperwriting tasks between the sub-activities.

Encouraged by the positive results ofthe iterations, we are working to test theactivity with more children and institu-tions. In sharing the lessons we learnedduring the design iterations we hope toinspire and support colleagues workingin the field... and get useful feedback!

Link:

[L1]https://www.epfl.ch/labs/chili/index-html/research/cellulo/

References:

[1] T. Asselborn, A. Guneysu, et al.“Bringing letters to life:handwriting with haptic-enabledtangible robots”, in Proc. of the17th ACM Conference onInteraction Design and Children(IDC '18). ACM, New York, NY,USA, 219-230, 2018.

[2] Ö.Özgür et al.: “Cellulo: Versatilehandheld robots for education”, in12th ACM/IEEE InternationalConference on Human-RobotInteraction, pp. 119-127. IEEE,2017.

[3] A. Özgür et al.: “DeclarativePhysicomimetics for TangibleSwarm Application Development”,Proc. of the 11th InternationalConference on Swarm Intelligence,ANTS 2018. Vol. 11172. No.CONF. 2018.

Please contact:

Arzu Guneysu OzgurEPFL, CHILI (Computer-HumanInteraction in Learning and InstructionLaboratory), [email protected]

Figure�1:�Robot-assisted�writing�activity�with�Cellulo:�Children�interact�with�the�haptic-enabled�Cellulo�robots�on�maps�displaying

the�grapheme�of�a�cursive�letter,�where�the�start�and�end�of�the�letter�are�indicated�with�a�starting�cue�and�a�trophy�respectively.�After

three�different�sub-activities�where�the�robots�provide�a�range�of�sensory�information,�a�final�team�activity�in�the�form�of�a�guessing

game�is�played.

“Computational thinking (CT) is goingto be needed everywhere. And doing itwell is going to be a key to success inalmost all future careers.” The words ofStephen Wolfram [L1] capture theurgency of the efforts to introduce CTinto educational curricula before highschool. In a parallel effort, robots arebeing increasingly used in educationalsettings around the globe, with someattempts to use robots to help studentsadvance CT skills. However, findingpedagogical designs that can helpdevelop such skills is quite challenging.

To address the current need, the“JUSThink” project [L2], started inSeptember 2018 at the Computer-Human Interaction in Learning andInstruction (CHILI) Laboratory at ÉcolePolytechnique Fédérale de Lausanne(EPFL), aims to improve the computa-tional thinking skills of children by exer-cising abstract reasoning with andthrough graphs, where graphs are posedas a way to represent, reason about andsolve a problem. JUSThink builds onexperience from earlier work [1] andaims to foster children’s understandingof abstract graphs through a collabora-tive problem-solving task, in a setupconsisting of a QTrobot and touchscreens as input devices (shown inFigure 1). JUSThink is being developedas part of EU’s Horizon 2020 ANI-MATAS Innovative Training Networkthat aims to improve the social interac-tion capabilities of robots for learningand education [L3].

In this activity, the design is scaffoldedtowards and inspired by pedagogical andlearning theories of collaboration andconstructivism whereas the role of therobot (an informed CEO, who still needshelp) is partially inspired by the protégéeffect. Furthermore, the design of theactivity is motivated to surface cues rel-evant to (i) learners’ engagement withthe task at hand, their partner and the

robot, and (ii) mutual understandingand misunderstandings between thelearners. Given the importance of theaforementioned cues towards the peda-gogical goal of the activity, the ultimategoal of JUSThink is to enable the designof a robot capable of understanding,monitoring and, if necessary, inter-vening.

For this purpose, we brought the setupto multiple international schools inSwitzerland over a span of two weeks,

where around 100 children (age: M =10.5, SD = 1.36; median = 10) partici-pated in pairs to have a one hour inter-action with the setup.

Children in pairs are welcomed to asetup as in Figure 1 by the QTrobot asthe CEO of a gold-mining companylooking to hire. After the robot’s wel-come, the children introduce them-selves to the robot and individuallysolve a brief test. Then, the robot tells

them the goal of the activity, which is tobuild railroads connecting all the com-pany’s gold mines, distributed in themountains of Switzerland, by spendingas little money as possible as the com-pany has a limited budget (hence,solving a “Minimum Spanning Tree”problem), and illustrates how they caninteract with the setup. After the signal“Are you ready? Let’s go!”, the childrenare assigned two roles, with one of themdrawing and erasing tracks while theother reasons about the moves based on

the costs shown, and the children swapthe roles every two moves. Once theyagree on a solution, children submit it tothe robot CEO, that, in addition togiving them basic guidance and encour-agement, tells them how close theirsolution is to the best possible solution.The collaborative turn-taking design,with a barrier between the children,each of whom always only having par-tial information, is inspired by twohypotheses: (i) partial information scaf-



“You Tell, I do, and We Swap until we Connect

All the Gold Mines!”

by Jauwairia Nasir, Utku Norman, Barbara Bruno and Pierre Dillenbourg (EPFL)

The Project JUSThink has a double goal: (i), to help train the Computational Thinking skills of children

with a collaborative, robot-mediated activity, (ii), to acquire insights about how children detect and

solve misunderstandings, and what keeps them engaged with a task, the partner or a robot. The

result? An abstract reasoning task with a few pedagogical tricks and a basic “robot CEO” that can

keep 100 ten-year-olds engaged, and, in turns, frustrated and jubilant!

Figure�1:�QTrobot�welcomes�children�to�the�activity.

folds towards collaboration when thereis a shared goal, and (ii) justifying pastand future moves verbally toself/partner based on cost can lead to aninitial grounding for abstract reasoning.

The children, surprisingly, withoutexception, reported really enjoying theactivity and expressed interest in inte-grating it within school activities eventhough they thought it was challenging.In addition to this, they really liked theirrobot CEO and found it to be knowl-edgeable, helpful and friendly.

For us as researchers, what now? Tohelp improve the learning outcomes inthis context of human-human-robotinteraction, this user study and themulti-modal data collection is just the

beginning. We aim to extract relevantbehaviours from the data generated(logs from the apps, facial and lateralvideos, audio) and relate it to learning toexplore models of engagement andmutual modelling in collaboration withSorbonne University and Télécom Parisin France. Eventually, the generatedmodels would lead to adapting the robotbehaviour effectively in real time toadvance learning as well as have appro-priate human-human-robot interactionin educational contexts so that the nexttime we go to schools, we are one stepcloser to “Educational Technologies forthe Future”.

Moral of the story: Children are moreaccepting of challenging educationalactivities than we think.

Links:

[L1] https://kwz.me/hEM[L2] https://kwz.me/hER[L3] http://www.animatas.eu/

References:

[1] J. Nasir, U. Norman, et al.: “RobotAnalytics: What Do Human-RobotInteraction Traces Tell Us AboutLearning?,” IEEE RO-MAN 2019Conference, 2019.

Please contact:

Jauwairia NasirEPFL, [email protected]

Utku NormanEPFL, [email protected]


Ullmer and Ishii describe tangible userinterfaces as giving “physical form todigital information, employing physicalartifacts both as ‘representations’ and‘controls’ for computational media” [1].The “Gestures in Tangible UserInterfaces” (GETUI) project [L1], cou-pled 2D and 3D gesture analysis withtangible user interfaces (TUIs) with theaim of achieving technology-basedassessment of collaborative problem-solving.

We ran exploratory user studies at twosecondary schools in Luxembourg andone in Belgium with 66 pupils. We usedthe multiuser interactive displayMultiTaction MT550 and a Kinect v.2.0depth sense camera to record 3D datarelating to the participants. The partici-pants’ task, presented as a microworldscenario on the MultiTaction device,was to build a power grid by placing androtating tangible objects or widgets,which represented industrial facilities(coal-fired power plants, wind parks,and solar parks) that produce electricity(Figure 1). The design and developmentof the microworld scenario was donethrough the COllaborative Problem

Solving Environment (COPSE), whichis a novel and unique software frame-work for instantiating microworlds ascollaborative problem-solving activitieson tangible tabletop interfaces [2].

We used the Kinect camera to explorethe behaviour of the participants duringthe collaborative problem solving. Themain problem we experienced withKinect was user identification. As iscommon in multi-user environments,users moved frequently in order toexplore different parameters on theTUI, with the result that their initial IDswere lost or exchanged, leading to mis-interpretation of the logging data. Also,the lightening and position of the Kinecthad to be selected carefully. An addi-tional technical limitation is recognitionof finger-based gestures, includingemblems (substitutes for words) andadaptors (gestures without consciousawareness used to manage our feel-ings). Another research-related draw-back is the definition of a gesture andparticularly of cooperative gestures.When exactly does a gesture start andwhen does it end? A gesture usuallypasses through up to five phases: prepa-

ration, prestroke hold, the stroke itself,poststroke hold, and retraction [3]. Inour TUI setting, the most prominentgesture type is pointing. What if, duringthe poststroke hold of a user A, user B isin the preparation phase of her own ges-ture? We consider cooperative gestureas a gesture sequence when two or moregestures, the first of which is alwayspointing, are performed simultaneouslyor consecutively by multiple users (notby the same user). However, the impactof the cooperative gesture has to beannotated manually, since it can be pos-itive, negative, or none.

The 2D gestures, i.e., gestures per-formed on the tabletop, are logged bythe COPSE software. We decided todevelop an application to link theCOPSE with Kinect, so that all per-formed gestures, both 2D and 3D arelogged. Therefore, our application con-sists of two components: i) the ClientReader (COPSE), which reads andtransfers information about the objectID and the TUI’s coordinates of theobjects, and ii) the Body Reader, whichreceives these coordinates and convertsthem into Kinect coordinates by using a

Gestures in Tangible user Interfaces

by Dimitra Anastasiou and Eric Ras (Luxembourg Institute of Science and Technology)

2D gestures have been extensively examined on surface and interactive tabletop computing, both in

the context of training of predefined datasets and “in the wild”. The same cannot be said for 3D

gestures, however. The current literature does not address what is happening above the tabletop

interfaces, nor does it address the semantics.

transformation matrix (see Figure 2).This matrix is created by the calibrationprocedure, where the TUI location andits plane are transformed into the Kinectcoordinates system. More informationis available in a technical report [L2].

The analysis and evaluation of gesturesis important both economically andsocially, with many fields using gestureas input or output in their applicationsfor instance, telecommunications,entertainment, and healthcare.Nonverbal behaviour is of particularimportance in collaborative and virtualenvironments. Until now, few studieshave addressed the nonverbal cuespeople display in collaborative virtualenvironments.

With the GETUI project we examinedcorrelations between 3D gestures andcollaborative problem-solving perform-ance using a TUI. We compared twogroups (high-achievers vs. low-achievers) and found that the pointinggestures were almost equal among thetwo groups (M = 25.7 for low-achieversand M = 25.6 for high-achievers), whilethe adaptors (head/mouth scratching,nail biting) were used slightly more fre-quently by the low-achievers, whereasthe emblems (thumbs up, victory sign)were used largely by high-achievers.We addressed and measured collabora-tion as one of the transversal skills ofthe 21st Century. TUIs and the visuali-sation of microworld scenarios can beused both for formal school education

and assessment as well as for vocationaltraining and modern workplacelearning. Today, in group settings, it isnot only the group problem-solving per-formance that matters, but also softskills, which include personal, social,and methodical competences. In thefuture, we plan to apply the assets ofGETUI to collaborative virtual environ-ments, in order to create and assessavatars’ nonverbal behaviour.

Links:

[L1] https://kwz.me/hKK[L2] https://kwz.me/hKS

References:

[1] B. Ullmer, H. Ishii, H.: “EmergingFrameworks for Tangible UserInterfaces”, IBM Syst. J. 39, 915-931, 2000

[2] V. Maquil, et al.: “COPSE: Rapidlyinstantiating problem solvingactivities based on tangibletabletop interfaces”, in Proc. of theACM on Human-ComputerInteraction 1(1), 6, 2017

[3] D. McNeill: “Hand and mind:What gestures reveal aboutthought” Chicago: University ofChicago Press, 1992.

Please contact:

Dimitra AnastasiouLuxembourg Institute of Science andTechnology, [email protected]



Figure�2:�Application�linking�the�recognition�of�2D�and�3D�gestures.

Figure�1:�Microworld�scenario�on�the�TUI.�


Computational thinking skills are con-sidered fundamental for our future dailylives. One approach to foster these skillsis to introduce primary school childrento programming and electronics throughplatforms that are easy to use, engagingand fun. Kniwwelino® [L1][1] is a plat-form that teaches computationalthinking in a way that is attractive andinteresting to children, while beingaccessible and easy to use.

Kniwwelino is a microcontroller-basedlearning environment consisting of a setof printed circuit boards (PCBs), anArduino Library and a visual program-ming interface. The main PCB, theKniwwelino board, proposes a 5×5LEDs matrix, an RGB LED and twopush buttons (see Figure 1). The micro-controller has Wi-Fi connectivity,enabling several boards to connect overthe internet. Additional PCBs can be

used as extensions and include differentsensors and actuators, such as an LED, abuzzer, a potentiometer, an ultrasonicdistance sensor, a touch button, or atemperature sensor. They can be con-nected to the main board via alligatorclips or other types of wires.

To program the hardware, a KniwwelinoLibrary [L2] is available as well as avisual programming interface [L3] (seeFigure 2) accessible via the webbrowser. In this interface, pieces of codeare represented by blocks and users canassemble their programs using drag anddrop, like the pieces of a puzzle.

The firmware is distributed to theKniwwelino board wirelessly viaWiFi. This has the advantage that noinstallation of drivers or software isrequired and any type of device(laptop, desktop and tablet) and oper-ating system can be used.

Kniwwelino provides a creative andhands-on approach to programming andelectronics. Children can rebuild, cus-tomise and invent a variety of projectscombining technological with non-tech-nological components. Example proj-ects include a wristband to send mes-sages and icons to friends or a weatherstation that shows current weather con-ditions available on the internet.

Kniwwelino was developed in an itera-tive, user-centric design approach,building upon results from previouswork and taking in account currenttechnologies available on the market.Throughout its development,Kniwwelino was tested with more than2,000 children who attended work-shops organised at different fairs.Participants provided feedback via ashort questionnaire.

The Kniwwelino workshops wereequally attended by girls and boys, themajority of whom had never codedbefore. Most participants found the

Kniwwelino: A Microcontroller-based Platform

Introducing Children to Programing and Electronics

by Valérie Maquil and Christian Moll (Luxembourg Institute of Science and Technology)

The Luxembourg Institute of Science and Technology (LIST) has developed Kniwwelino, a microcontroller-

based development platform that uses a visual programming interface to help children discover

programming and electronics in a creative and hands-on approach.

RGB LED

3.3V Output Ground

Button A

Battery

LED Matrix

Accessory Connector (I2C)

Digital In-Outputs

Convertisseur Analogue/Digital

Button B

Figure�1:�The�Kniwwelino�board�contains�a�5×5�LEDs�matrix,�an�RGB�LED,�two�push�buttons

and�several�pins�for�connecting�additional�hardware.

Figure�2:�In�the�visual�programming�interface,�users�can�assemble�their�programs�using�drag

and�drop,�like�the�pieces�of�a�puzzle.

Kniwwelino workshops very fun andthe included tasks relatively easy. Theywould like to use Kniwwelino again andto try other programming tools.

Currently, LIST is collaborating withseven primary and secondary schoolsin Luxembourg to evaluate the feasi-bility of using Kniwwelino in theLuxembourg’s classrooms. Every pilotschool was provided with a classroomkit containing in total 15 pupil boxes,each including a Kniwwelino board, 14different extensions, and cables to con-nect them. They also received a hand-book with pedagogical activities, proj-ects to build and creative material tosupport pupils in developing new proj-ects. Over the coming months, teachers

will test these materials in their class-rooms and document their impressionsand experiences.

Kniwwelino was developed by theLuxembourg Institute of Science andTechnology (LIST) in collaborationwith the National Youth Service(Service National de la Jeunesse –SNJ), a government organisation incharge of non-formal education inLuxembourg, as well as SCRIPT, agovernment organisation responsiblefor formal education. The project issupported by the Luxembourg NationalResearch Fund (FNR) under the PSPFlagship (2016-2019) and Jump POC(2019-2020) schemes.

Links:

[L1] https://www.kniwwelino.lu/[L2] https://github.com/kniwwelino[L3] https://code.kniwwelino.lu

Reference:

[1] V. Maquil, et al.: “Kniwwelino: ALightweight and WiFi EnabledPrototyping Platform for Children”,in Proc. of the 12th InternationalConference on Tangible, Embedded,and Embodied Interaction, pp. 94-100, ACM, 2018.

Please contact:

Valérie Maquil Luxembourg Institute of Science andTechnology, [email protected]



Working in emergency medicine is char-acterised by constraints that constitute ahigh-risk constellation: need for situa-tional assessment and decision-makingas well as initiation of appropriate emer-gency measures under time pressure,often under adverse external conditions,and, at the same time, with little or nofault tolerance. Virtual realities (VR) areincreasingly used as a simulation tech-nology in emergency medicine educa-tion and training, in particular for thetraining of “non-technical” skills (clin-ical and procedural reasoning) and team-work skills. Experimental studies havedemonstrated that VR is equivalent, oreven superior, to traditional trainingmedia (i.e., patient actors or manikins)for training purposes.

In the project EPICSAVE, a highlyimmersive room-scaled multi-user 3Dvirtual simulation environment wasdeveloped for medical training sce-narios. This project involved an interdis-ciplinary consortium incorporatingexpertise from all relevant disciplines,i.e., paramedic training academies, med-ical and media education, media design

research, and production of virtual 3Dlearning environments. Following atwo-year interdisciplinary and iterativedevelopment and evaluation process,the second prototype of the virtual sim-ulation environment now consists ofdifferent virtual emergency locationswith integrated virtual patients of dif-ferent age groups [L1]. The virtual envi-ronment contains more than 60 virtualobjects, e.g., interactively usable virtualinstruments for the diagnosis andtherapy of the virtual patient:• Change of posture and undressing of

the virtual patient• Monitor-based measurement, e.g.,

electrocardiogram, blood pressure • Clinical examination, e.g., breathing

sounds• Medical treatment, e.g., administra-

tion of oxygen, application of infu-sions.

The hardware equipment for a teamtraining with two trainees consists oftwo VR head-mounted displays (HTCVive®), four input devices (HTCController®), and two computers withdisplay, which are used by the trainers

to control the simulation and follow thetraining. The trainees and their interac-tions are transmitted to the VR in realtime by means of a motion-trackingsystem. In the VR, the trainees are rep-resented by avatars. This enables col-laborative activities connecting the realworld with the VR (e. g., handing overmedical equipment). The VR systemautomatically records the training ses-sions and documents important diag-nostic steps or therapeutic interventions[1]. The virtual patient represents a mul-titude of pathological parameters andsymptom characteristics: • Psychomotor condition• Different states of consciousness • Breathing sounds• Pulse frequency and strength• Cyanosis, blue coloration of the skin.

This goes beyond the current capabili-ties of commercially available, costlyhigh-fidelity simulators or patientactors. The EPICSAVE virtual simula-tion environment enables the realistictreatment of complex, dynamic, and rareemergencies without compromising realpatients (see Figure 1).

virtual Simulation Environment

for Medical Training

by Thomas Luiz, Dieter Lerner, Dominik Schnier (Fraunhofer IESE)

Virtual Realities (VR) are increasingly used as a simulation technology in emergency medicine

education and training. In the project EPICSAVE, a highly immersive room-scaled multi-user 3D

virtual simulation environment was developed for medical training scenarios. This enables a

realistic and sustainable training experience.


The highly immersive VR has madepossible a dynamically changeable, real-istic, and three-dimensional visualisa-tion of different clinical environmentsand patients from different perspectives[2]. This enables spatial positioning ofthe trainees in the virtual 3D emergencyenvironment and fosters the feeling ofbeing present in this scenario. A highlevel of presence experience correlatespositively with variables for learning ortraining effectiveness. Furthermore, ahigh degree of realism increases thelikelihood that the learning experienceswill be transferred to real environments.As part of the project, the virtual simula-tion environment was evaluated withfive study groups. The following param-eters were measured during eachtraining session: presence experience,usability, cyber sickness, assessment oftraining effectiveness, cognitive load,current learning motivation, and up-to-date knowledge of the respective emer-gency scenario. The participants of thefive studies rated the VR simulation

training above average in terms oftraining effectiveness, quality of trainingexecution, and potential for improvingmotivation and immersion [3].

The EPICSAVE VR will continue to bedeveloped even after the project hasended. In the follow-up BMBF projectViTAWiN (Virtuell-augmentiertesTraining für die Aus- und Weiterbildungin der interprofessionellen Notfallver-sorgung [Virtual-augmented training foreducation in interprofessional emer-gency care], FKZ 01PV18006), mixed-reality technologies are now being com-bined with VR. The ViTAWiN approachpursues the goal of expanding VRthrough the integration of haptic inputand output devices, such as augmentedvirtuality (AV). Thus, the previouspotential of training and learning in VR(non-technical skills) can be extended tothe training of technical skills. In addi-tion to this realistic and three-dimen-sional visualisation, a perceptualisationof virtual objects (virtual patient, virtual

equipment) can also be achieved. Thelearning- and training-relevant contentsin the VR will be presented to thetrainees using means of visual, auditory,and haptic-tactile interaction [L2].

The project EPICSAVE (EnhancedParamedIC vocational training withSerious games And VirtualEnvironments) was funded by theGerman Federal Ministry of Educationand Research (BMBF) and theEuropean Social Fund of the EuropeanUnion (EFS) (duration: 03/2016-02/2019, FKZ 01PD15004).

Links:

[L1] https://www.epicsave.de [L2] https://www.vitawin.info

References:

[1] J. Schild, et al.: “EPICSAVE –Enhancing Vocational Training forParamedics with Multi-user VirtualReality”, in IEEE SeGAH, pp. 1-18, 2018. DOI:10.1109/SeGAH.2018.8401353.

[2] D. Lerner, T. Luiz: “Close toreality. Learning with virtualpatients” [Nah an der Realität.Lernen mit virtuellen Patienten],Intensiv 27 (2), 64–69, 2019. DOI:10.1055/a-0821-3183

[3] D. Lerner, D. Wichmann, K.Wegener: “Virtual reality simula-tion training in emergency para-medic training” [Virtual-Reality-Simulationstraining in derNotfallsanitäterausbildung], retten!2019; 8(04), 234–237. DOI:10.1055/a-0820-8614

Please contact:

Dieter Lerner Fraunhofer IESE, GermanyResearch Program [email protected]

Figure�1:�The�virtual

patient�in�the�virtual

simulation

environment,�(source:

TriCAT�GmbH,�Ulm).

Mixed Reality and Wearables in Industrial Training

by Ralf Klamma (RWTH Aachen University), István Koren (RWTH Aachen University) and Matthias Jarke(Fraunhofer FIT and RWTH Aachen University)

The Horizon 2020 project WEKIT has created an industrial training platform, combining sensor

technologies, artificial intelligence and mixed reality. The platform creates training materials on the fly

and delivers them in a standardised manner.

With digitisation of workplacesimpacting many professions, industrialtraining needs to adapt accordingly.

Wearables show enormous potential tohelp in this space, and we envisagethree major shifts taking place. First,

industrial training often includestraining in manual labour processes.Quantitative digital ethnography based

on big data from wearable devicescould foster an understanding of theseprocesses in digital workplaces.Second, once the big data are acquired,automated analysis will be needed. Toachieve this, artificial intelligence andmachine learning will enter industrialtraining. Finally, the interface betweenhumans and machines will move awayfrom flat computer screens to moreimmersive forms of feedback in artifi-cial spaces in mixed reality environ-ments.

The European H2020 project WEKIT(Wearable Experience for KnowledgeIntensive Training, 2015-2019) aims tointroduce wearables into industrialtraining scenarios. The projectaddresses three scenarios, displayed inFigure 1: ground training for astronauts,training for aircraft maintenance per-sonnel in arctic rescue missions andtraining of medical personnel in the useof 4D ultrasound diagnostic devices.

From a technical perspective, the mainoutcome of the project is theWEKIT.ONE platform [1]. The plat-form consists of a self-developed hard-ware component: a bank of sensors in aself-designed vest with an electronicboard (PCB), connected to a MSHoloLens and more sensors in mobilephones and arm wrists.

The hardware is connected to a recorderand a player. The recorder enables thecreation of learning materials byrecording the sensor data of expertsexecuting procedures for the trainingscenarios. The data are stored in a data-base and analysed by the WEKIT.ONEsoftware. The player recognises thetraining situation either automaticallyor as instructed by a trainer.Subsequently, the player supports thetraining through a number of environ-mental augmentations, such as virtualtraces on the ground leading from onestation to the next, or by “ghost hands”

indicating the manipulation of deviceslike the cutting of a wire through the useof an augmented head-mounted device.

As a pedagogical innovation, WEKIThas developed a new instructionaldesign model. Tasks, support informa-tion, procedural (manual) knowledgeand practices have been categorised foruse within the platform. In this context,WEKIT has been addressing the ques-tion of how we can describe, store,retrieve and exchange training sce-narios together with necessary contex-tual information in a standardised way.An IEEE-SA working group has begunthe standardisation of our AugmentedReality Learning Experience Model(ARLEM). An open-source modeleditor is available at [L1]. Public projectdeliverables are available on the web-site [L2] and a start-up companyWEKIT ECS (Experience CapturingSolutions) [L3] is exploiting the results.

This new field of learning with wear-able technologies is thoroughlyexplored in a new book published bySpringer [2]. The interface betweenhumans and machines moves awayfrom traditional cognitive tools, just aswe moved from typewriters to digitalthinking tools. Humans are increasinglyinteracting with machines, and tech-nology is increasingly being used intraining scenarios [3]. Data fusion andartificial intelligence can bring togetherdata from heterogeneous sources,helping us to recognise human activitiesin both human-human and human-robotcollaboration scenarios. By making theanalytical results available in mixedreality spaces, training, analytics andinterventions can take place within thesame space without the media breaksand loss of context that occur in tradi-tional training.

New forms of sensor data fusion andprocessing, including big data visualanalytics, are enabling innovative

research into the changing workplace.In conducting this research, data protec-tion, privacy, ethical considerations andadherence to workplace legislationmust have the highest priority. Like anynew technology, these tools have thepotential to be misused, and a broadsocial acceptance is key to their success.

Links:

[L1] https://kwz.me/hEA[L2] http://wekit.eu/[L3] https://wekit-ecs.com

References:

[1] B. Limbu et al.: “WEKIT.One: ASensor-Based Augmented RealitySystem for Experience Capture andRe-enactment,” in LNCS,Transforming Learning withMeaningful Technologies, M.Scheffel, et al., Eds., Springer,2019, pp. 158–171.

[2] I. Buchem, R. Klamma, F. Wild:“Perspectives on WearableEnhanced Learning (WELL):Current Trends, Research, andPractice”. Cham, Switzerland:Springer Nature Switzerland AG,2019.

[3] R. Klamma, R. Ali, and I. Koren:“Immersive Community Analyticsfor Wearable Enhanced Learning,”in LNCS, Learning andCollaboration Technologies.Ubiquitous and VirtualEnvironments for Learning andCollaboration, P. Zaphiris and A.Ioannou, Eds., Springer, 2019, pp.162–174.

Please contact:

Ralf Klamma RWTH Aachen University, [email protected]



Figure�1:�Scenarios�from�the�WEKIT�project�©�WEKIT�and�Mikhail�Fominykh.


Mixed-reality allows the merging ofphysical and virtual worlds, creatingenvironments where physical and digitalobjects co-exist and interact in real-time.It has been used in areas ranging fromentertainment and health applications tomilitary training. Extended reality (XR)is defined as a form of mixed realityenvironment that comes from the fusionof ubiquitous sensor/actuator networksand shared online virtual worlds, toencompass all the possibilities of realitywarping technology. Statista reportedthat social media (i.e., Facebook) had1.49 billion monthly active users in thesecond quarter of 2015. Imagine edu-tainment environments so engaging thatpeople could stay in them for hours andnever tire of them! This is where XRcomes into the picture.

The role of future XR interactivesystems in edutainment As an emerging technology, virtual,augmented and mixed reality (XR) notonly supplements the dynamic notion ofthe instructional practices but alsoincorporates sensory modalities, suchas, touch, sight and hearing. Severalstudies reveal the potential benefits of

XR in educational contexts, includingimproving the user’s level of achieve-ment, motivation, knowledge retentionand engagement. This technology canalso guide targeted behaviour change toimprove the way that various activitiesare undertaken so that the user beginsto take the desired actions in differentcontexts whilst deriving more enjoy-ment from their tasks [3]. XR technolo-gies are breathing life into the notionthat edutainment can be accomplishedanywhere; not just within the confinesof a classroom. XR is undoubtedlypoised to change the way users deliverand acquire information, knowledge,and skills, in playful learning environ-ments [1]. This is bolstered by the risein popularity of VR and AR devices inEurope, with sales in Western Europereaching an impressive 283 thousandunits of VR/AR head-mounted displaysin the second quarter of 2016.According to a recent report publishedby Allied Market Research, the globalmixed reality market was valued at$123.2 million in 2017, and is projectedto reach at $5,362.1 million by 2024,growing at a CAGR of 71.6% from2018 to 2024.

XR technologies in educationMotivation to learn is a complexprocess, and what motivates on studentmay not motivate another. Recentstudies have shown that personalisededutainment systems improve motiva-tion: students respond favourably toflexible or reduced training/studyinghours, course adjustments, or part-timestudying and autonomy – all of whichare primary areas of intervention forpersonalised systems [2]. However, wedon’t currently know what role thesesystems play in motivating students tospend more hours acquiring an under-standing of complex concepts using lab-oratory equipment where layouts andmodalities are more potent. While mul-tisensorial collaborative edutainmentsystems hold great potential to improvethe interventional aspects of education,very few such systems are used to teachscientific concepts. This means that cur-rent interventional “devices” (e.g. vir-tual assistants, physics based XR inter-actions and cognitive aware visual ana-lytics) in personalised systems couldhave both positive and negative educa-tional effects, which, at best, reducestheir potential impact (e.g., to induce

Personalized Interactive Edutainment in Extended

Reality (XR) Laboratories

by Aris S. Lalos (ISI, ATHENA R.C.), Chairi Kiourt (ILSP, ATHENA R.C.), Dimitrios Kalles (Hellenic Open University) and Athanasios Kalogeras (ISI, ATHENA R.C.)

XR-LAB envisages developing a highly innovative & interactive extended reality (XR) platform that will empower

users, including educators and trainees, to create easily accessible and sustainable edutainment experiences.

This will be achieved using holographic interfaces and gamified elements that will develop gradually as familiarity

with interactive features is gained. The proposed approach is expected to create convenient, safe, economic,

rapid, flexible and user-friendly spaced educational tools that challenge, engage and prepare students for their

real scientific experiment in remote STEM education laboratories.

Figure�1:�XRLabs�conceptual�architecture.� Figure�2:�XRLabs�cutting-edge�technologies.�

Figure�3:�Interaction�with�VR�environment

with�motion�sensing�controllers.

real knowledge acquisition) and, atworst, might lead to rebound effects(e.g. ironic effects, poor understandingdue to information overload or over-trust). If these personalised edutainmentsystems are to be effective, theyundoubtedly require smart incentivemechanisms that offer higher qualityexperiences, motivating users to engagein creative activities.

The XRLab caseThe XRLab is a platform focusing onmodern education based on student-centred educational systems withdynamic content, which is not affectedby geographical, financial and timeconstraints, in the way that conven-tional educational methods are. Theproposed platform focuses on labora-tory training for all levels of education.Within this framework, the systemsunder development are based on cut-ting-edge technologies, such as interac-tive virtual, augmented and mixedreality environments. The XRLab plat-form combines three main researchareas of Semantic Web, namely: knowl-edge management in education, gamifi-cation and educational systems. Figure1 depicts the main architecture of theXRLab platform. According the mainstructure graph, all users (authenticatedor not) have basic access (limited) to thesystems of the platform. Authenticatedusers, such as students and teachers, areassociated with their status and aregiven additional services. The main

structure of the XRLab includes cut-ting-edge technologies/methodologies,shown also in Figure 2, focusing on thedevelopment of a user-friendly educa-tional environment. An example of theXRLab platform equipped with motionsensing controllers is shown in Figure 3,where students interact with the dynamicobject of a virtual laboratory. XRLab fol-lows a user-centred design approach,designing and evaluating the projectdevelopments and the overall, integratedXRLab system in realistic conditions inHellenic Open University. Our mainobjective during the evaluation phase isto investigate whether students who usethe XRLab as a supplement to traditionallearning methods learn scientific con-cepts and experimental skills more easilyor if they are cognitively overloaded bythe large amount of information, themultiple technological devices, and thecomplex tasks they have to deal with.The impact of such a holistic approach ishuge and the foundations laid here areexpected to result in a widespread adop-tion of sensor-based XR interaction plat-forms in a wider range of edutainmentcontexts.

This work is supported by the projectXRLAB - Virtual laboratories usinginteractive technologies in virtual,mixed and augmented reality environ-ments (MIS 5038608) implementedunder the Action for the StrategicDevelopment on the Research andTechnological Sector, co-financed by

national funds through the Operationalprogram of Western Greece 2014-2020and European Union funds (EuropeanRegional Development Fund).

Link: http://XRLab.eu/

References:

[1] I. Radu: “Why should my studentsuse AR? A comparative review ofthe educational impacts ofaugmented-reality”, in Proc. ofIEEE International Symposium onMixed and Augmented Reality(ISMAR) (pp. 313–314), 2012.

[2] E. Paxinou, et al.: “A 3D virtualreality laboratory as asupplementary educationalpreparation tool for a biologycourse”, European J. of Open,Distance and E-Learning, 2018

[3] K. Gardelis, A. Lalos, K.Moustakas: “Development of anEco-Driving Simulation TrainingSystem with Natural and HapticInteraction in Virtual RealityEnvironments”, in Proc. of the 13thInt. Joint Conf. on ComputerVision, Imaging and ComputerGraphics Theory and Applications- Vol. 2: HUCAPP, pp. 94-101,2018. ISBN 978-989-758-288-2

Please contact:

Aris S. LalosISI, Athena Research Centre, [email protected]



The iMuSciCA project [L1] (InteractiveMusic Science Collaborative Activities)has made a first step towards integratingmusic into STEM (Science, Technology,Engineering and Mathematics) educa-tion. iMuSciCA not only brings arts intothe heart of STEM, following the para-digm of STEA(rts)M education, but italso combines pedagogy with innova-

tive technologies. The new methodintroduced by this project has beendesigned, applied, tested and, whenevernecessary, re-adjusted in a real educa-tion setting. To this end, iMuSciCAinvolved a multi-disciplinary teamcomprising academics, artists and tech-nology experts who co-developedlesson plans alongside an innovative set

of interactive STEAM educational toolson a unified workbench [L2].

A novel educational approachSTEAM is an educational paradigm thatis gaining ground in the European edu-cational system and iMuSciCA isamong the first projects to examine in-depth the implications of STEAM in

Music in Education through Technology

by Maximos Kaliakatsos-Papakostas, Kosmas Kritsis, Vassilis Katsouros (Institute for Language andSpeech Processing, Athena Research and Innovation Centre)

Combining Science, Technology, Engineering, Arts and Mathematics (STEAM) into a unified pedagogical

framework, enables students to directly identify connections between abstract concepts and elements

of the real-world. The iMuSciCA project has made the first steps towards developing tools, methods and

lesson plans for stimulating creativity in learning and providing the basis for deeper learning. This is

achieved with cross-disciplinary lesson plans implemented through the iMuSciCA workbench: an

innovative suite of web-based software tools around STEAM.


deeper learning. Until recently, techno-logical education and the arts have beenconsidered as separate subjects, how-ever, bridging the two fields leads to across-fertilisation process that benefitslearning in both fields. According to theiMuSciCA approach, immense educa-tional dynamics can develop by incor-porating creativity, expressed inmusical expression and experimenta-tion, in STEM learning environments.In iMuSciCA this is achieved through aweb-based workbench [1] that inte-grates advanced core enabling tech-nologies, including 3D design andprinting of musical instruments, bodytracking sensors for gesture recognition,interactive pens and tablets as well assound generation and processing tools.

Various tools have been developedbased on the aforementioned technolo-gies that allow users/students to realiseand develop connections between andacross concepts in music and STEM.This helps students creatively discoverlaws of physics[2], relations in mathe-matics and phenomena in engineeringand technology. This is achieved byallowing students to move from onefield to another with different startingpoints, practically by changing educa-tional environments inside the work-bench, implementing the inquiry-basedscience education (IBSE) phases, i.e.,engage, imagine, create, analyse, com-municate and reflect. Some lesson planshave been developed in the context ofthe iMuSciCA project, but the work-bench also enables teachers to be morecreative and expressive in their teachingby authoring their own scenarios, lever-aging the combined dynamics and theopenness of the workbench as well asthe ease with which users can jump fromone STEAM environment to the other.

Expertise and technologies: a cross-discipline consortiumThis multi-facet approach to learningcould only be achieved by a consortiumcomprising partners who not only bringunique expertise but are also in a posi-tion to understand and develop ideasand technologies across disciplines.ATHENA Research and InnovationCentre focuses on interaction with 3Dtechnologies, but also has expertise indeveloping educational as well as musicsoftware. The University College atLeuven-Limburg (UCLL) developed theoverall pedagogical design and evalua-tion of the outcome of the entire project.

Ellinogermaniki Agogi, a private schoolin Greece, implemented pilot tests andhelped in developing lesson plans,involving a multidisciplinary team ofteachers (from music to physics). TheInstitute de Recherche et deCoordination Acoustique Musique(IRCAM) in France developed the real-time sound synthesis algorithm and actu-ation technology, involving tech expertswho are also musicians. Leopoly, aHungarian SME, contributed with theirexpertise in 3D interactive object design.Cabrilog from France offered tools forinteractive education in mathematics.Spanish WIRIS contributed with theirhandwriting recognition technologiesalong with musical tools that took advan-tage of mathematical formulas and geo-metric shapes. The University ofFribourg in Switzerland worked onhandwriting analysis and workbenchintegration and, with ATHENA, indeploying biometrics sensors forassessing implications of the iMuSciCAapproach in (deeper) learning.

Evaluating the iMuSciCA approachResearch teams in Greece, Belgium andFrance initially investigated theusability of the iMuSciCA workbenchby co-creating lesson plans and educa-tional scenarios with teachers and thenconducted pilot studies betweenNovember and December 2017. Bothteachers and students gave positivefeedback along with some commentsthat were integrated in subsequent ver-sions of the workbench. The final ver-sion of the workbench during the timespan of the project, included completelocalisation with translations on alltools; this, along with the improvedusability, allowed for a final pilot testing

in two phases from September 2018 toMay 2019, to assess the effect of thisapproach on deeper learning, withoutthe barrier of the UI being presented in aforeign (to non-English speakers) lan-guage. The deeper learning evaluationprocedure established by the pedagog-ical team is based on an approach pro-posed by the Hewlett Foundation and itwas based on questionnaires for studentsand teachers. The results of this study(currently available on [L2]) indicatedthat the iMuSciCA approach is effectivefor facilitating deeper learning.Acceptance of iMuSciCA by teachersseems promising, with teachers’ com-munities forming in the countriesinvolved in the pilot studies.

Links:

[L1] http://www.imuscica.eu/[L2] https://workbench.imuscica.eu/

References:

[1] K. Kritsis, et al.: “iMuSciCA: AWeb Platform for ScienceEducation Through MusicActivities”, in Proc. of the WebAudio Conference 2019 (WAC’19). Trondheim, 2019.

[2] E. Andreotti, R. Frans: “Theconnection between physics,engineering and music as anexample of STEAM education”,Physics Education, 54, 2009.http://doi.org/10.5281/zenodo.3358408

Please contact:

Maximos Kaliakatsos-Papakostas,Vassilis Katsouros, Athena Researchand Innovation Centre, Greece [email protected],[email protected]

Figure�1:�The�iMuSciCA�cross-disciplinary�approach,�among�many�other�activities,�allows�the

design�of�virtual�3D�virtual�instruments�using�real-world�physical�quantities,�along�with

informative�visualisations�of�sound�(timbre/harmonics)�and�music.

The relationship between dance educa-tion and new technologies is still imma-ture, but evolution in fields such asmotion capture, augmented, mixed andvirtual reality, combined with data, infor-mation and knowledge representationand analysis, have great potential to ben-efit the dance education of the future [1].

The WhoLoDancE (Whole-bodyInteraction Learning for DanceEducation) project’s main goal is todesign and apply innovative digital edu-cation technologies that will benefit thecommunity of dance practitioners,experts, researchers, students, educators,choreographers, as well as the widercommunity of amateurs and profes-sionals and the interested public.

The project’s objectives can be sum-marised as: (i) to investigate bodilyknowledge by applying technologiessuch as computational models to auto-matically extract characteristics that arerelevant to both the quantitative andexpressive characteristic of movement;(ii) to preserve cultural heritage by cre-ating a motion capture repository ofdance motions, with built-in methodsallowing enrichment through annota-tion, segmentation and synthesis, (iii) todevelop innovative ways of teachingdance, such as via life-sized visualisa-tion of the body through differentavatars in mixed reality - e.g., usingHoloLens and sonification of move-ment; (iv) to revolutionise choreog-raphy, by designing and organising aninteractive repository of motion capture,providing choreographers and danceteachers a powerful tool to blend and getinspired by an infinite number of dancecompositions from different dance tradi-tions; (v) to extend access to and prac-tice of dance, by providing access to thecreated dance database through com-mercially available low-end motion cap-ture devices like the MS Kinect, low-cost sensors and wearables.

The project focused on four dancegenres that represent different forms of

intangible heritage and / or performingart: ballet, contemporary, flamenco andGreek folk. Each of these genres alsopresents unique requirements related tomovement vocabulary and teachingapproach.

A major outcome of the project is thecreation of a big database of over 780curated dance movement sequences thathave been recorded using motion cap-ture technologies. A conceptual frame-work [2] has been proposed to guide the

recording process and develop the userscenarios and requirements for the finalwhole-body interaction experiences.The resulting toolkit of WhoLoDancEincludes a set of web-based tools toanalyse, segment, annotate and blendmovement as well as interactive experi-ences that integrate augmented andmixed reality, sonification of movementand visualisation of the human bodyand movement in different avatars andenvironments [3]. In addition, severalevaluation events and performativeworkshops have been held throughoutEurope (UK, Greece, Italy, Spain) tar-

geting dance educational institutionsand wider audience.

WhoLoDancE project, as a collabora-tion between technology experts,researchers, dance educators, choreog-raphers and artists, used four dancegenres as use-cases and a proof-of-con-cept to open a dialogue on how dancemovement can be captured and trans-mitted to the next generations. In fact,both the outcomes and open issues thathave been reported by the project, sug-

gest that there is still considerablefuture work to be done to design theideal digital educational experience fordance. The process is complicated bythe complexity and richness of humanmovement as well as the diversity andcultural pluralism that is integrated intodance. For example, teaching of Greekfolk dance, as a traditional dance formand an important part of cultural her-itage, includes much more than theteaching of the movement itself. Everydance is tightly connected with rhythm,lyrics, the location of its origin, the cus-toms, the climate, the costume, the con-



dance Education and digital Technologies

by Katerina El Raheb and Yannis Ioannidis (“Athena” Research and Innovation Center and Nationaland Kapodistrian University of Athens)

WhoLoDancE (Whole-body Interaction Learning for Dance Education), a three-year EU funded project,

has developed innovative technologies for dance practice, creation and education, focusing on four

dance genres: ballet, contemporary, flamenco, and Greek folk.

Figure�1:�Two�Greek�Folk�dancers�from�Lykeion�ton�Hellinidon�(The�Lyceum�Club�of�Greek

Women)�recorded�using�optical�Motion�Capture�at�Motek�Entertainment�Studio,�for�the�creation

of�WhoLoDancE�Movement�Library.

text in which the particular dance is per-formed and by whom (e.g., elders, men,women), the stories and traditions.

There are huge technical challengesassociated with analysing movement,and to address it properly requires a col-laborative effort between multiple com-puter science fields, including motioncapture, signal processing, data ana-lytics, semantic technologies, artificialintelligence, human computer interac-tion, graphics and digital artistic per-spectives. The interdisciplinary chal-lenges have proven even greater, and inmost cases the design process of ameaningful digital pedagogical tool is acomplex constraint satisfactionproblem, with conflicting objectives(e.g., making tools that are reusable andscalable vs. identifying and respectingthe needs of the particular dance tradi-tions).

Upon completion of the project, most ofthe partners, including “Athena”

Research and Innovation team will becontinuing research into how to makerelevant technologies more usable, cost-effective and accessible to a wider audi-ence. As a research team, we will keepinvestigating effective and meaningfulways to document, represent andtransmit dance movement and knowl-edge.

WhoLoDancE (Whole-body InteractionLearning for Dance Education) is aclosed three-year Research andInnovation Action funded under theEuropean Union’s Horizon 2020Programme that investigates the designand application of digital tools to dancelearning and education [L1].WhoLoDancE is a collaborationbetween European technology research,educational and intangible cultural her-itage institutions with creative industryand dance companies.

Link:

[L1] http://www.wholodance.eu/

References:

[1] K. El Raheb, et al.: “DanceInteractive Learning Systems: AStudy on Interaction Workflow andTeaching Approaches”, ACMComputing Surveys (CSUR) 52.3(2019): 50.

[2] A. Camurri, et al.: “WhoLoDancE:towards a methodology forselecting motion capture dataacross different dance learningpractice”, in Proc.of the 3rdInternational Symposium onMovement and Computing. ACM,2016.

[3] A. Rizzo, et al.: “WhoLoDancE:Whole-body Interaction Learningfor Dance Education”, CIRA@EuroMed. 2018.

Please contact:

Katerina El Raheb, National and Kapodistrian Universityof Athens, Greece, [email protected], +302107275188


Acknowledging the omnipotent impactof technology on the domain of educa-tion, we have pursued the vision of a stu-dent-oriented and educator-friendly“Intelligent Classroom” that supportsstudents in their journey to acquiringknowledge. The classroom simulationspace [R1] located within the AmIFacility [L1] provides an ideal testbedfor assessing the effects of intelligenttechnologies on key aspects of the edu-cational process. Designing and devel-oping ambient applications that shapethe “Intelligent Classroom” is anongoing activity and an open researchendeavour, integrating and assessingemerging technologies that introducenew educational opportunities and inter-action paradigms.

More specifically, the hardware infra-structure includes both commercial andcustom-made artefacts, which are

embedded in traditional classroomequipment and furniture, while variousambient facilities are also available formonitoring the overall environment andthe learners' activities. One of the keyclassroom artefacts, the “Smart StudentDesk”, has been drastically re-designedsince its first realisation in 2010, whenit featured a vision-based back-pro-jected multi-touch screen. Now, thedesk artefact features a more appealingand ergonomic design, embedding a 24-inch wide, all-in-one computer, securedin a rotatable steel frame. At the sametime, the next version of the studentdesk is already under development, fea-turing a modular design, where cus-tomisable surfaces can be added orremoved on demand, in order to supportthe specific needs of each course. Thisnew generation desk will furtherenhance students’ engagement andmotivation, offering hands-on experi-

ence and providing personal studyspaces with specialised equipment.

The design process of the classroomboard followed a course similar to thatof the students’ desk. The first versionof the “Intelligent Classroom Board”featured a commercially available inter-active white board, while the secondversion used a 70-inch, 4K TV alongwith a multi-touch panel in order toenable collaboration and increase theinteractive area. Following the recentadvances in projection technologies,which can transform plain walls intointeractive displays, the current versionof the classroom board extends beyonddigital devices; two out of the four wallsof the “Intelligent Classroom” (i.e., thewall in front of the entire class and oneside wall) act as interactive smartboards, where educational content (e.g.,multimedia, notes, exercises) can be

Intelligent Classroom: Materialising the vision

of Ambient Intelligence for Education

by Asterios Leonidis, Maria Korozi, Margherita Antona and Constantine Stephanidis (FORTH-ICS)

The ICS-FORTH Ambient Intelligence (AmI) Programme is a long-term horizontal interdisciplinary RTD

Programme aiming to develop pioneering human-centric intelligent technologies and environments that

seamlessly support everyday human activities and enhance well-being through human-technology symbiosis.

presented, while they are also used tocreate attractive, situational (i.e.,course-, topic- and discussion-specific)environments, where students can beimmersed into an “environment” (e.g., acave or a rainforest). Finally, the designof the “Teacher Workstation” hasshifted away from the traditional deskand now enables educators to monitorand manipulate every aspect of the“Intelligent Classroom” (e.g., ambientfacilities, educational software, intelli-gent behaviour, automations) from acomfortable technologically-aug-mented armchair.

The underlying software of the“Intelligent Classroom” has evolvedgradually over the years and has nowreached a satisfactory level of maturity,creating a unified environment that pro-motes teaching and learning. In partic-ular, the ClassMATE and the AmI-Solertis [R2] frameworks support fun-damental services, such as interoper-ability of heterogeneous intelligentservices, synchronous and event-basedcommunication, resilience, security,etc. Additionally, they empower contex-tual awareness, thus enabling the class-room to immediately respond andorchestrate the available intelligentartefacts (e.g., boards, desks, walls) inorder to effectively and efficientlyaddress the needs of students and edu-cators. Furthermore, such infrastructure

features an adaptive content retrievalmechanism and a learners’ behaviourknowledge library, to enable the intelli-gent environment to personalise theeducational content to each student’sactual learning needs. Additionally, theclassroom features LECTOR [R3], aneducator-oriented platform that helpsidentify behaviours that require reme-dial actions (e.g., student is not payingattention) and deliver appropriate inter-ventions when students and teachersneed assistance (e.g., a motivational cueto encourage participation, a suggestionto adjust the current pedagogicalapproach in order to re-engage stu-dents). As far as the end-user applica-tions are concerned, CognitOS is a web-based window manager that hosts edu-cational applications and instantiates acommon “Look-n-Feel” across the var-ious classroom artefacts, thus trans-forming the classroom into a unifiedenvironment, rather than a group of iso-lated units. On top of that, CognitOSfeatures mechanisms that enable the ini-tiation of interventions dictated byLECTOR in order to: (i) attract the edu-cator’s attention in problematic situa-tions, and (ii) re-engage distracted orunmotivated students in the educationalprocess.

Up until now, many ambient applica-tions (e.g., teaching Greek as a foreignlanguage, independent living skills



Figure�1:�3D

representation

of�the

“Intelligent

Classroom”

displaying�the

“Smart�Student

Desk”,�the

“Intelligent

Classroom

Board”�and�the

“Teacher

Workstation”.

Figure�2:�Snapshots�of�CognitOS�dashboards�and�multimedia�application�running�on�the�“Smart�Student�Desk”.

training for children with cognitive dis-abilities) and educational games (e.g.,geography game, spelling game) havebeen implemented for the “IntelligentClassroom”. These applications havebeen tested through formal evaluationexperiments, while dozens of demon-strations offered to visitors of the“Intelligent Classroom” have given theopportunity for informal evaluations.Our future plans include multiple, in-vitro, summative, user-based evaluationexperiments with educators and stu-dents, in order to assess the effective-ness of each system, as well as the class-room environment as a whole.

Link:

[L1] http://ami.ics.forth.gr

References:

[1] M Korozi, et al.: “ Shaping theIntelligent Classroom of theFuture”. In C. Stephanidis & M.Antona (Eds.), HCI International2019 – Late Breaking Posters, Vol.37 of the combined Proceedings ofHCI International 2019, Orlando,FL, USA, y (pp. 200-212).Springer, 2019

[2] A. Leonidis, et al.: “The AmI-Solertis System: Creating UserExperiences in SmartEnvironments”, in Proc. ofPerCAM 17, in the context of theWiMob 2017, Rome, Italy (pp.151-158), 2017

[3] M. Korozi, et al.: “Lector: TowardsReengaging Students in theEducational Process inside SmartClassrooms”, in P. Horain, et al.(Eds.), Intelligent HumanComputer Interaction, Proc. ofIHCI 2017, Evry, France (pp. 137-149). Springer, LNCS 10688, 2017

Please contact:

Constantine StephanidisFORTH-ICS, Greece+30 2810 [email protected]


Testing in education has traditionallybeen a process of checking and evalu-ating students’ knowledge - generallywith exams. Testing can also serve as alearning tool, however, with informationbeing retrieved but not recorded. Testingis known to have long term learningbenefits, although there is no consensusabout the mechanism by which the“testing effect” operates. Some studiesassume that retrieving a particular pieceof information from our memory alsoactivates notions associated with it [1].Others suggest that recalling strengthensthe memory imprint corresponding to agiven piece of information [2].

Regular testing also encourages studentsto learn continuously instead of “cram-ming”, giving them frequent feedback

about their knowledge and progress,resulting in a less failure-avoidant atti-tude towards practising and reducedanxiety about testing. Today’s Z- andAlpha-generation students, who havegrown up in the digital age, are habitu-ated to using information technologyand are motivated by online tools thatensure instantaneous feedback.

The benefits of the testing-effect havebeen widely discussed via pilot-studiesduring foreign language learning [3],and it is utilised in popular learningapplications like Duolingo and Quizlet.Its effectiveness hasn’t yet been investi-gated in mathematics, however, which

relies on conceptual ways of thinkingand abstract concepts. Furthermore, it isextremely difficult to test theoreticalhypotheses and the results of pilotstudies in real-life using conventionaltools, as it is almost impossible to carryout experiments over a large sample(hundreds of students) and over a longerperiod (e.g., a whole semester).Consequently, a novel tool is required tomonitor students’ learning behaviourand testing performance.

In this study, we focused on a first-yearundergraduate mathematics course(Calculus 1) for engineers at BudapestUniversity of Technology andEconomics. Calculus 1 is one of the mostimportant basic subjects in the curriculaof engineering programmes, in which

shared (or continuous) learning is essen-tial for deep knowledge on which the fur-ther engineering subjects are grounded.However, it is common among studentsto study in a campaign-like way, so thatthey can succeed in the course withoutacquiring profound knowledge.Additionally, freshmen arrive at univer-sity with very different mathematicalbackground. Therefore, it is essential toprovide other opportunities, in additionto lectures and seminars, to catch up, forself-development and to provide motiva-tion for continuous learning.

To provide these opportunities and tomonitor learning performance, we used

EduBase Classroom [L1], a cloud-based educational platform developedby the members and students of ourmathematics methodology group (theworkflow of our methodology is illus-trated in Figure 1). EduBase is a deviceand platform-independent learning-management system (LMS) providing acustomisable teaching and testing inter-face that covers a wide spectrum ofexaminations (e.g., homework, tests,and exams), which can be managed viaa digital classroom. The test system ofEduBase was developed for mathemat-ical testing in which parametrised tasks,function evaluations, vector and matrixformulations are also implemented. Thequizzing system also provides contin-uous data on learning performance,enabling the tutors to continuouslymonitor learning performance andprogress during the semester, which isnot possible with traditional methods ofteaching. The EduBase Test System isproving to be an excellent tool not onlyfor online education and motivation butalso for scientific research in the fieldsof cognitive sciences and teachingmethodology.

After analysing the results at the end ofthe semester in a class of 115 freshmen,we can conclude that the novel teachingmethodology based on continuousretrieval-based learning in EduBase hasfulfilled our expectations. The compar-ison with the control group from pre-vious years, who learnt the conven-tional way, has shown that drop-outrates have fallen dramatically, and morestudents are aiming for better grades. Inweeks in which new topics were intro-duced, the time students spent onlineincreased significantly compared toother weeks. In this phase of learningstudents used testing for learning andtraining themselves for solving prob-lems rather than checking their acquiredknowledge. The end-of-semester resultsconfirmed that scores achieved during

Innovative Monitoring of Learning Habits and

Motivation in undergraduate Mathematics Education

by Brigitta Szilágyi (Budapest University of Technology and Economics), Szabolcs Berezvai (Budapest University ofTechnology and Economics) and Daniel Horvath (EduBase Online Ltd.)

The Educational platform EduBase ensures continuous testing opportunities and analyses the learning behaviour

and performance of undergraduate engineering students in a mathematics course. Compared to the conventional

method, students using EduBase were found to be more motivated to practice consistently throughout the

semester and acquire deeper level of knowledge.

Figure�1:�The�workflow�of�teaching�methodology�using�EduBase�Test�System�for�online�testing,

motivation�and�detailed�analysis�of�student�data.

this phase of learning are not neces-sarily indicative of grades at the end;high marks can be achieved by studentswho struggle with the homework butpractise a lot.

A detailed statistical analysis ofEduBase’s data showed that studentswho practised consistently achievedsignificantly better results at the examand acquired a deeper level of knowl-edge than those who crammed for tests.A survey of students confirmed thatonline testing not only supported theirlearning but also motivated them tocomplete maths exercises frequently.

To conclude, we propose a novel test-based methodology that can be appliedin real-life university classes withexcellent efficiency. EduBase Test

System ensures not only that studentspractise consistently and can takeadvantage of catch-up opportunities,but also helps the tutors to continuouslymonitor learning habits and perform-ance throughout the semester. In furtherresearch, this tool could also be appliedto analyse relationship between dailyroutine and performance, and the phe-nomenon of forgetting and re-learning.

Link:

[L1] https://www.edubase.net/

References:

[1] S. K. Carpenter, H. Pashler, N. J.Cepeda: “Using tests to enhance8th grade students’ retention ofU.S. history facts”, AppliedCognitive Psychology, 23, 760–771,2019. doi:10.1002/acp.1507

[2] J. G. W. Raaijmakers, R. M.Shiffrin: “SAM: A theory ofprobabilistic search of associativememory”, in G. Bower (Ed.), Thepsychology of learning andmotivation (Vol. 14), New York:Academic Press, 1980.

[3] A. Keresztes, et al.: “Testingpromotes long-term learning viastabilizing activation patterns in alarge network of brain areas”,Celebral Cortex 24 (11): 3025-3035, Oxford University Press,2014.

Please contact:

Brigitta Szilágyi, Szabolcs BerezvaiBudapest University of Technologyand Economics, [email protected],[email protected]



When a Master of Sciences on EdTech

becomes an International Community

by Margarida Romero, Saint-Clair Lefèvre (UCA, INSPE, LINE) and Thierry Viéville (Inria)

Nelson Mandela said, “Education is the most powerful weapon which you can use to change the

world”. How do we best achieve this? Part of the answer may lie in SmartEdTech, a Master of

Science (MSc) program specialising in digital education.

The MSc SmartEdTech program [L1] atUniversity Côte d’Azur aims to developa quality approach to digital educationwith a co-creative and participative edu-cation. The program is embedded withina socioeconomic sensibility, specificallyrelated to the Global Goals of the UnitedNations in relation to education [L2].The challenge is to both improve today'seducation and create tomorrow’s educa-tion, by means of breakthrough digitalpedagogies.

As illustrated in Figure 1, the end goal isto allow everyone to develop the “21stcentury skills”, focusing in criticalthinking, creativity, problem solving,collaboration and computationalthinking as key skills for citizens [1]. Wealso illustrate how this could apply tocomputational thinking as a key compe-tency to solve problems through theknowledge of formal systems (how tocode) and physical systems (what is asensor, how a network works, etc.).Through computational thinking devel-opment, the students should be able tonot only use technology for education

but engage in the critical approach ofanalysing problem situations, identifytechnologies to solve these problems.

The MSc SmartEdTech has beendesigned with an objective of inclu-sivity to ensure that the students areselected based on their projects andpotential [2]. In order to develop aninternational community, 90% of theprogram is online, based on openreusable resources in English and somein French. This blended training alsoincludes two intensive weeks, to reallymeet each other, and to learn and shareand intern periods, inside either anEdTech company, or an educationalresearch lab, while some studentsdevelop their own spin-off company.Msc SmartEdTech students areexpected to become the future e-learning or digital innovation projectmanagers, instructional designers,teacher or e-learning instructors,edTech consultants or researchers in thelearning sciences. Students within theMSc SmartEdTech program are profes-sionals from a diverse range of back-

grounds including quality assuranceprofessionnals in Higher Education,teachers and university professors, appdevelopers and freelancers in the fieldof EdTech. They are also engaged inteaching activities within the MSc,embodying the “communities of prac-tices” (CoP) paradigm, in whicheveryone has the potential to act as aresource for the community.

As students, they will address the chal-lenges of designing and integratingEdTech in their different educationalcontexts. Our conviction is that theyhave to learn how to learn, concretely,via effective competence development.The activities developed in the programsupport active learning approachesthrough co-creation approaches, Digitalgame based learning (DGBL), makereducation and maker culture. With thesedifferent activities, students areengaged to develop a critical under-standing of digital sciences andembrace a paradigm of (co)creators oftechnologies and, therefore, are morethan just EdTech consumers. The core


of our paradigm is project-basedlearning (PBL), meaning that ourapproach puts everyone's professionalproject at the heart of learning, rightfrom the start. As such, all coursesbecome an opportunity to realise agiven project, while each student hasthe opportunity to engage in smallgroup teamwork or receive specificcoaching.

This MSc benefits from a threefold sci-entific approach, led by the Learning,INnovation and Education (LINE)research lab [L3], in computer scienceand education science.

This EdTech MSc not only teacheshigh-tech tools, but also low-techapproaches. For instance, in order toteach children how a computer works,“unplugged activities”, coupled withcreative programming initiation andeducation robotics, are the pillars ofcomputational thinking learning [3]. Ina nutshell, the main point is to use an

everyday situation as a metaphor for atheoretical computer science concept.

Our approach is agnostic with respect tothe available tools. This positioning iscrucial given the huge social and envi-ronmental challenges facing our planet.Whilst the use of digital tools, in educa-tion or elsewhere, is neither a panacea,nor the inverse, there are a few things ofwhich we are certain: The core chal-lenge is pedagogy: can digital tools helpus improve the pedagogy we want todevelop?

The MSc has been running for 18months and to date our team of nineteachers from six countries have helpedabout 40 colleagues to contribute to“changing the world through educa-tion”. Out of the 12 pioneer studentstwo have created a company or otherstructure, three are working in researchpositions, two have been hired byEdTech companies, and the remainderare working on personal projects.

Perhaps the most interesting and moti-vating aspect of this experience is theway in which this tiny, lively, globalgroup has started to form a multicul-tural, interdisciplinary and internationalcommunity, with a common objective,sharing ideas and practices, being (seeFigure 2). This is exactly what this ini-tiative is striving for: to go beyondstructured teaching and learning and tobuild an international community ofpractice (CoP) in 21st century educa-tion.

Links:

[L1] https://kwz.me/hK4 [L2] https://kwz.me/hK0[L3] http://unice.fr/laboratoires/line

References:

[1] M. Romero, A. Lepage, B. Lille:“Computational thinking develop-ment through creative program-ming in higher education”,International Journal ofEducational Technology in HigherEducation, 14(1), 42, 2017.

[2] M. Ciussi, et al.: “Numérique etéducation … vous avez dit#CreaSmartEdtech ?”, 2018.binaire.lemonde.fr,https://tinyurl.com/y6n5pwvr

[3] D.Menon, et al.: “Going beyonddigital literacy to developcomputational thinking in K-12education”, L. Daniela, SmartPedagogy of Digital Learning,Taylor & Francis (Routledge),9780367333799, 2019.https://hal.inria.fr/hal-0228103

Please contact:

Margarida RomeroUCA, INSPE, LINE, [email protected]

Figure�1:�The�21st�century�educational�skills

includes,�beyond�knowledge�and�know-how,

important�personal�and�collective

competences,�and�societal�values.

Figure�2:�The�#CreaSmartEdtech�Msc�international�map,�featuring�40�students�coming�from�25

countries�gathered�to�learn�together,�and�start�a�truly�international�community.

The�MSc�SmartTech�is�developped�in�partnership�with�EducAzur.

EducAzur�[L1]�is�a�French�cluster�currently�comprising�22�public

and�private�institutions�[L2],�including�Inria,�ERCIM�and

University�Côte�d'Azur.�Its�mission�is�to�bring�together�stakeholders

in�order�to�encourage�innovative�projects�between�local�and

national�ecosystems�and�to�accelerate�experimentation�in�school,

higher�education�and�professional�training.

Examples�of�innovative�projects�include:�the�EducAzur�and�Class'Code�partnership�to

facilitate�learning�code�in�the�Provence�Alpes�Côte�d’Azur�(PACA)�region;�the�organization

of�conferences�on�education�and�artificial�intelligence;�and�its�contribution�to�the�Valbonne

city's�digital�educational�excellence�program.�Together�with�other�French�clusters�and�with

the�support�of�the�Banque�des�Territoires,�EducAzur�has�contributed��to�the�creation�of�the

Federation�of�French-speaking�EdTech�clusters.

[L1]��https://educazur.fr/�[L2]�https://educazur.fr/members/

EducAzur

Technology-enhanced learning (TEL),also known as educational technology, isa research field that explores new waysof learning enabled by technology anddesigning new technologies that cansupport learning in new ways. The termalso refers to a learning activity when itis supported by technology in practice.Such practical application of TEL,referred to as the digitisation of educa-tion, is being implemented on all levelsacross Europe. However, digitisationtrends are diverse, rapidly changing, andoften lack a clear application idea whenit comes to putting the theory into prac-tice - thus, educators still employ tradi-tional forms of teaching. Thorough plan-ning is needed to implement these hugetechnological changes, requiring profes-sionals in TEL with expertise in peda-gogy, the design and use of technologiesfor learning and strategic expertise inpolicy making and change managementat institutional and societal levels.

Schools, universities, private and publicorganisations, and industry are allexpressing a need for interdisciplinaryexpertise in this area. At a societal level,there is also a growing understandingthat in an increasingly complex worldlearning at all ages must become stan-dard practice.

There is a current lack of TEL expertswho combine knowledge from all rele-vant perspectives, who can strategicallydrive and operationally implement digi-tisation of learning, envision innovativetechnologies for learning, and rigorouslycreate evidence about the effectivenessand efficiency of technologies. Suchexpertise is generally acquired at thedoctoral level of education.

The Doctoral Education for Technology-Enhanced Learning (DE-TEL) projectaims to address the challenges of digiti-sation of education in Europe and thecurrent lack of expertise by improvingdoctoral education in the area of TEL.

DE-TEL brought together nine interna-tionally renowned universities and theEuropean Association of Technology-Enhanced Learning (EATEL), coordi-nated by an ERCIM member, theNorwegian University of Science andTechnology.

The goal of DE-TEL is to bring doctoraleducation in TEL to a new level withhigh-quality resources and a new inter-nationally designed program to supportbetter curricular integration and avoidfragmentation of the digitisation agendain Europe. The project grounds thedesign of the new program in the bestpractices in TEL doctoral educationacross Europe as well as institutionaland national requirements.

DE-TEL will create a new opportunityin an area of research that has practicalapplication across Europe as the digiti-

sation of education is increasinglyrequired to solve multiple challenges.Digitisation requires pedagogicalexpertise in using and technologicalexpertise in designing technologies forlearning, and strategic expertise inpolicy making and change manage-ment. However, stakeholders currentlylack combined knowledge from allthree perspectives, on how to strategi-cally drive and operationally implementdigitisation of learning and training,envision innovative technologies forlearning, and rigorous methods to createevidence for the effectiveness and effi-

ciency of technologies as part of socio-technical interventions. Such expertiseis required at all levels and at all typesof organisations. There is also agrowing, recognised understanding byall relevant stakeholders, from EUpolicy makers to local schools, thattechnologies have the potential to pro-vide ubiquitous access to learning mate-rials, and an engaging, personalised andscalable learning experience.

Doctoral education in TEL aims to helpdoctoral candidates to develop suchexpertise. It provides knowledge frommultiple relevant perspectives andallows empirically-based decisions tobe made regarding implementing TELsolutions in practice. The goal of DE-TEL is to bring doctoral education inTEL to a new level with high-qualityresources and a new internationallydesigned program to support better cur-

ricular integration and avoid fragmenta-tion of the digitisation agenda inEurope.

The DE-TEL initiative is brought for-ward by an established community ofEuropean researchers and educatorsgravitating around the activities of theEATEL (The European Conference onTechnology-Enhanced Learning -ECTEL and The Joint Summer Schoolon Technology-Enhanced Learning -JTELSS). The consortium will reflecttheir expertise in doctoral educationinto a new internationally validated pro-



dE-TEL - A European initiative for doctoral

Education in Technology-Enhanced Learning

by Mikhail Fominykh and Ekaterina Prasolova-Førland (NTNU)

The Doctoral Education for Technology-Enhanced Learning (DE-TEL) project is developing high quality

resources and an internationally recognised program to train PhD students in the area of educational

technology. The interdisciplinary expertise that this program will foster is essential for an efficient and

effective digitalisation of education in Europe.

Figure�1:�DE-TEL�consortium.�


gram in TEL. The new program will beextended with rich and professionallyproduced open educational resources.

All resources produced in the projectwill be open and available free ofcharge to any interested parties.Moreover, the project is working on anonline infrastructure that will allowboth the program and the resources tobe discussed, updated, translated and

extended by the community during andafter the project.

The project will also organise two openface-to-face training events dedicated topiloting the new program and educa-tional resources, preliminary scheduledfor 2021 and 2022.

DE-TEL has received funding from theEuropean Union’s Erasmus Plus pro-

gram for the period 2019-2022 undergrant agreement number 2019-1-NO01-KA203-060280.

Link:

https://ea-tel.eu/de-tel/

Please contact:

Mikhail Fominykh, NTNU, Norway(+47) [email protected]

smart classrooms encompass a largerange of different products, we candefine them as physical environments inwhich teaching–learning activities arecarried out using computerised devicesand various techniques and tools: signalprocessing, robotics, artificial intelli-

Ethical Teaching Analytics in a Context-Aware

Classroom: A Manifesto

by Romain Laurent (Univ. Grenoble Alpes/LaRAC), Dominique Vaufreydaz (Univ. Grenoble Alpes, CNRS,Inria, Grenoble INP, LIG), and Philippe Dessus (Univ. Grenoble Alpes/LaRAC)

Should Big Teacher be watching you? The Teaching Lab project at Grenoble Alpes University proposes

recommendations for designing smart classrooms with ethical considerations taken into account.

Figure�1:�Setting�of�the�Context-Aware�Classroom�designed�by�the�Teaching�Lab�Project�at�University�Grenoble�Alpes.�

The smart education market is flour-ishing, with a compound annual growthrate regularly announced in doubledigits (+31% between 2014 and 2018,+18% between 2019 and 2025). Amongthese EdTech investments, smarts class-rooms that combine AI and adaptivelearning are in the spotlight. Although

gence, sensors (e.g., cameras, micro-phones) and effectors (e.g., loud-speakers, displays, robots).

Despite all the hype around smart class-rooms, it seems that the ethical and pri-vacy aspects of these developments areseldom considered. This issue is not

incidental: it conditions not only theclassroom atmosphere and the ethics ofthe institution, but also the quality ofinterpersonal relationships within theclassroom and teaching–learningprocesses — which universities oftenclaim to support.

This leads to a series of questions.Which signals will be captured andstored? How will the consent of theindividuals be gathered? Should themachine monitor everything that hap-pens in the classroom? Will unduly col-lected signals be excluded (e.g., acamera-captured personal message, amicrophone-captured conversation, aninappropriate behaviour)? If yes, how?Ex ante (by the machine) or ex post (bythe operators)? Will the machine remainspectator or will it act in real time usingeffectors? How will the system thence-forth react to learners’ and eventeachers’ behaviours that might be con-sidered dilatory? What outcomes are welikely to see from the highly delicateinterpretation of features inferred fromdata? In response and to counter anysurveillance, what opacifying behav-iours or layers will be implemented bythe interactors? What trust can then beestablished between the teacherenhanced (and possibly monitored) bytechnology, the student under scrutinyand the institution receiving thisdataveillance [1]?

Teaching and learning are fragile activi-ties that rely above all on the commit-ment of participants. This mutual com-mitment is rooted in trust and respect,which can only flourish and prosper onthe integrity of the relationship betweenthe teachers, the students and the institu-tion that oversees it. This integrity pavesat last the way for relatedness, known ascritical in student’s achievement [2].The introduction of smart technologiesto the classroom, if not framed byexplicit ethical and privacy-compliantprinciples and purposes, could easily beused for dataveillance of its participants,jeopardizing classroom interactions andcorrupting the effectiveness ofteaching/learning processes [1].

We designed a smart classroom(Teaching Lab project at GrenobleAlpes University, funded by the “PIA 2IDEX formation program”), with essen-tial ethical safeguards in mind thatrespect individuals’ privacy as a prereq-uisite for the physical, technological

and theoretical development of thesespaces. The class smartness is exploitedin two directions. On the one hand, itauthorises the basic collection of educa-tional data (in particular the distributionof attention between teacher, studentsand learning materials), and their pro-cessing by machine learning tech-niques. On the other hand, it subordi-nates them to an ethical framework thatrespects the protagonists, following thesafeguards for privacy and data protec-tion in ambient intelligence [3]. To doso, we will use advanced machinelearning techniques to shed light onglobal features while obfuscating localones. Our objective is twofold. First, weowe privacy to each participant to theextent that everyone’s privacy is inter-woven with that of everyone else.Secondly, we aim to not bias the gen-uineness of observed behaviours. Weassume there that each individual’s pro-tection is a good way to strengthen theglobal gathered information (e.g. inter-actions, relatedness).

For this purpose, we propose four keyguidelines as the core of the TeachingLab design:

1. We develop an ethics-by-designapproach: beyond their common con-tribution in terms of signal process-ing, machine learning techniques willbe used to anonymise, at their source,attendees’ data and filter any itemlikely to alter the confidence of theprotagonists in the instructional flow.Thus, the Teaching Lab obfuscates allparticipants, while focusing on mean-ingful events and constructs (gaze,teacher cognitive load, hand, deictic,gesture, non-personal devices, voiceactivity and overlap, prosody).

2. We promote a global rather thanlocal methodology: students’ behav-iours will never be individuallytraceable; the machine learningbeing limited to restore a globalisedpicture of the occurrence of behav-iours. In other words, our datareports about the whole classroombut ignores individuals.

3. Finally, machine learning will beintentionally constrained to adelayed feedback, filtered by theresearch team, excluding any real-time monitoring or ad hominemreports. Our aim is only to supportinteractors’ possible professionaldevelopment (students included), bystudying and providing feedback

about the factors reinforcing orweakening attention in class.

4. The Teaching Lab will not make high-stakes decisions, monitor staff per-formance or student insulation. Fol-lowing the University of Edinburgh’sLearning Analytics’ principles andpurposes, we contend that data andalgorithms can contain and perpetuatebiases, and that they never provide thewhole picture about human capacityor likelihood of success.

These guidelines are intended to ensurethe privacy, security and trust of everyindividual within the smart classroom,thereby preserving relationships, inter-actions, and the research for improvingteaching and learning processes.

Links:

[L1] University of EdinburghPrinciples and purposes for LearningAnalytics: https://kwz.me/hK7[L2] Teaching Lab Project website:https://project.inria.fr/teachinglab/

References:

[1] R. Crooks: “Cat-and-Mouse games:Dataveillance and performativity inurban schools”, Surveillance andSociety, vol. 17, no. 3/4, pp.484–498, 2019.

[2] G. Hagenauer, S. E. Volet:“Teacher-student relationship atuniversity: an important yet under-researched field”, Oxford Reviewof Education, vol. 40, no. 3, pp.370–388, 2014.

[3] P. De Hert, S. Gutwirth, A.Moscibroda, D. Wright, and G.González Fuster, "Legal safeguardsfor privacy and data protection inambient intelligence," Personal andUbiquitous Computing, vol. 13, no.6, pp. 435–444, 2008.

Please contact:

Romain LaurentUniv. Grenoble Alpes/LaRAC, [email protected]

Dominique VaufreydazUniv. Grenoble Alpes, CNRS, Inria,Grenoble INP, [email protected]

Philippe DessusUniv. Grenoble Alpes/LaRAC, [email protected]



41

Eu

rop

ea

n R

ese

arc

h a

nd

In

no

va

tio

n

Research and Innovation

KANdINSKY Patterns:

A Swiss-Knife for the

Study of Explainable AI

by Andreas Holzinger (Medical University Graz andAlberta Machine Intelligence Institute Edmonton,Canada), Peter Kieseberg (University of AppliedSciences, St.Poelten) and Heimo Mülller (MedicalUniversity Graz)

Kandinsky Patterns are mathematically describable,

simple, self-contained, hence controllable test datasets

for the development, validation and training of

explainability in artificial intelligence (AI) and machine

learning (ML).

In machine learning we design, we develop, we test, and weevaluate algorithms, which can learn from data and extractknowledge and make predictions. This is very helpful fordecision support systems and decision making, particularlyin the medical domain. One family of machine learning algo-rithms is deep learning, which is now extremely successful.Deep learning reaches human-level performance in classifi-cation tasks even in the medical domain, with a classificationperformance of 92% and better. However, the big problemhere is its “black-box” nature, i.e., a medical expert is cur-rently unable to retrace, replicate and understand the under-lying explanatory factors of why the 92% have beenachieved, for instance. Particularly in medicine, the questionof why is often more important than the classification resultitself [1], therefore the field of “explainable AI” is becomingmore and more important [2].

Human experts can explain certain results very well,because humans are able to use abstract concepts. Causalityand concept learning are important to understand howhumans extract so much information, often from just a fewdata points, and to contrast this with machine learningwhich is now addressing “explainable AI”. For the study ofexplainable AI, we have designed and developed an experi-mental explanation environment for testing explainabilityconcepts of both human intelligence and artificial intelli-gence. We made this environment fully accessible for theinternational machine learning community and called itKandinsky Patterns, in memory of the great painter WassilyKandinsky. But there is another story behind these patterns:In the past we have observed how pathologists work [1].Pathology is a very interesting medical specialty; patholo-gists explain and interpret geometric objects and geometricstructures (Figure 1). They even speak of geometrical archi-tectures, and they observe and interpret shapes, objects,colour, similarity, Gestalt phenomena, etc., and finally comeup with an explanation in written form - the diagnosis. For anumber of reasons, this process is very difficult for machinelearning algorithms.

The Kandinsky Patterns enable the study of such phenomenaand to test, benchmark and evaluate machine learning algo-rithms under mathematically strictly controllable conditions[L1], [L2], but which at the same time are accessible and



understandable for human observers. It is also possible toproduce a ground truth. That means, we can produce imagesthat the tested classifier should evaluate as true, therebyallowing various ways of testing machine learning algo-rithms. Thus, Kandinsky Patterns can be used as a kind ofintelligence test for algorithms [3]. Kandinsky Patterns allowvalidation of results of arbitrary machine learning algo-rithms, but most of all, explainable AI methods can be tested,evaluated and further developed based on these experiments.This even allows for the development of completely newexplanation methods. Moreover, with our KandinskyGenerator we can produce “ false patterns” that can be usedto test the robustness of algorithms in a controlled setting.This will be extremely important in the future, as adversarialexamples have already demonstrated their potential inattacking security mechanisms applied in various domains,especially medical environments. Last, but not least,Kandinsky Patterns can be used to produce “counterfactuals”– the “what if”, which is difficult to handle for both humansand machines - but can provide new insights into the behav-iour of explanation methods.

In conclusion, Kandinsky Patterns (Figure 2) can be used as“IQ-Test for machines”, for causality research and to eval-uate explainable AI methods, i.e. to use causality as measure-ment [1].

Links:

[L1] https://human-centered.ai/project/kandinsky-patterns/ [L2] https://www.youtube.com/watch?v=UuiV0icAlRs

References:

[1] A. Holzinger, et al.: 2019. Causability and Explainabil-ity of Artificial Intelligence in Medicine. Wiley Inter-disciplinary Reviews: Data Mining and KnowledgeDiscovery, 9, (4), 2019, doi:10.1002/widm.1312

[2] R. Goebel, et al.:“Explainable AI: the new 42?, in Inter-national Cross-Domain Conference for Machine Learn-ing and Knowledge Extraction (pp. 295-303), 2018.doi: doi:10.1007/978-3-319-99740-7_21

[3] A. Holzinger, M. Kickmeier-Rust, H. Müller:“KANDINSKY Patterns as IQ-Test for Machine Learn-ing”, in International Cross-Domain Conference forMachine Learning and Knowledge Extraction (pp. 1-14). Doi:10.1007/978-3-030-29726-8_1

Please contact:

Andreas HolzingerexAI Lab, Alberta Machine Intelligence Institute,Edmonton, Canada and Medical University Graz, [email protected]

Figure�1:�A�typical�interpretation�task�in�pathology:�The

(human)�pathologist�interprets�and�explains�geometrical

architectures,�shapes,�objects,�colour,�similarity,�etc.,�and

produces�an�explanation�–�the�medical�report�[1].

Figure�2:�Kandinksy�Patterns�as�an

experimental�environment�for�explainable�AI�–

to�answer�the�question�of�why�the�classificatory

came�up�with�92%�[L2].

42


Figure�2:�Rebuild�under�clustered�placement.Figure�1:�Clustered�and�declustered�placement�of�codewords�of�length�three�on�six

devices.�X1,�X2,�X3�represents�a�codeword�(X�=�A,�B,�C,�.�.�.�,�L).

symbols of the affected codewords. When a clustered place-ment scheme is used, the lost symbols are reconstructeddirectly in a spare device, as shown in Figure 2.

When a declustered placement scheme is used, as shown inFigure 3, spare space is reserved on each device to tem-porarily store the reconstructed codeword symbols beforethey are transferred to a new replacement device. The rebuildbandwidth available on all surviving devices is used torebuild the lost symbols in parallel, which results in smallervulnerability windows compared to the clustered placementscheme.

The redundancy in the block-coding scheme and in the repli-cation over several storage units, adds to the cost of thesystem. Another cost is the transmission capacity requiredbetween the storage units for block recovery. An additionalperformance cost is incurred by the update operations associ-ated with the redundancy scheme.

Reliability is assessed via the Mean Time to Data Loss(MTTDL) and the Expected Annual Fraction of Data Loss(EAFDL) metrics. For such a storage system to operate effi-ciently with good quality of service, one is faced with amulti-dimensional optimisation problem. We have tackledthis problem theoretically by developing a method to derivethe MTTDL and EAFDL metrics analytically for variousdata placement schemes.

The methodology uses the direct path approximation anddoes not involve Markovian analysis [1]. It considers themost likely paths that lead to data loss, which are the shortestones. It turns out that this approach agrees with the principleencountered in the probability context expressed by thephrase “rare events occur in the most likely way”. The relia-bility level of systems composed of highly reliable compo-nents is essentially determined by the “main event”, which isthe shortest way to failure [2].

The analytical reliability expressions derived can be used toidentify redundancy and recovery schemes as well as dataplacement configurations that can achieve high reliability.Superior reliability is achieved by a distributed data place-ment scheme, which spreads redundant data associated withthe data stored on a given device in a declustered fashionacross several devices in the system.

Building Reliable Storage

Systems

by Ilias Iliadis (IBM Research – Zurich Laboratory)

A huge and rapidly growing amount of information needs

to be stored on storage devices and on the cloud. Cloud

users expect fast, error-free service. Cloud suppliers want

to achieve this high quality of service efficiently to

minimise cost. Redundancy, in particular erasure coding,

is widely used to enhance the reliability of storage

systems and protect data from device and media failures.

An efficient system design is key for optimising operation.

Today’s storage systems must store an enormous amount ofdata across multiple storage units. Regardless of the qualityof the storage technology, errors will occur. It is thereforepossible that one of these storage unit could fail in itsentirety. To further protect data from being lost and toimprove the reliability of the data storage system, replica-tion-based storage systems spread replicas corresponding todata stored on each storage device across several otherstorage devices. This is done in a way that allows the data tobe recovered even if a whole storage unit fails. To improvethe efficiency of the replication schemes, “erasure coding”schemes that provide a high data reliability as well as highstorage efficiency are deployed.

User data is divided into blocks (or symbols) of a fixed size(e.g., sector size of 512 bytes) and complemented with paritysymbols to form codewords. From these codewords, theoriginal data can be restored. Also, these codewords can beplaced on the available devices in different ways, whichaffects the reliability level achieved. The two basic place-ment schemes, clustered and declustered, are shown inFigure 1. The clustered placement scheme stores data and itsassociated parities in a set of devices with the number ofdevices being equal to the codeword length whereas thedeclustered placement spreads data and its associated paritiesacross a larger number of devices.

When storage devices fail, codewords lose some of theirsymbols, and this leads to a reduction in data redundancy.The system attempts to maintain its redundancy by recon-structing the lost codeword symbols using the surviving

Figure�3:�Rebuild�under

declustered�placement.

(a)�Normalized�MTTDL��(b)�Normalized�EAFDL

(a)�Spare�space�reserved�in�each�device��(b)�Distributed�rebuild

Figure�4:�Reliability�versus�codeword�length�for�various�storage�efficiency�values.

Links:

General information about cloud storage systems:https://www.zurich.ibm.com/cloudstorage/

References:

[1] I. Iliadis: “Reliability Evaluation of Erasure Coded Sys-tems under Rebuild Bandwidth Constraints”, Interna-tional Journal on Advances in Networks and Services,vol. 11, no. 3&4, pp. 113-142, December 2018.

[2] I. Iliadis and V. Venkatesan: “Most Probable Paths toData Loss: An Efficient Method for Reliability Evalua-tion of Data Storage Systems”, International Journal onAdvances in Systems and Measurements, issn 1942-261x, vol. 8, no. 3&4, pp. 178-200, December 2015.

[3] I. Iliadis, et al.: “Disk Scrubbing Versus IntradiskRedundancy for RAID Storage Systems”, ACM Trans.Storage, vol. 7, no. 2, article 5, 42 pages, July 2011.

Please contact:

Ilias IliadisIBM Research – Zurich [email protected]

Contrary to general assumption, we found that increasingcodeword length does not necessarily improve reliability.This is demonstrated in Figure 4, which plots reliability as afunction of codeword length (m) for various storage effi-ciency values. It demonstrates that increasing codewordlength initially improves reliability, but at some point relia-bility begins to degrade. This is due to two opposing effects:on the one hand, larger codeword lengths imply that code-words can tolerate more device failures, but on the otherhand, they result in a higher exposure degree to failure aseach of the codewords is spread across a larger number ofdevices. Further insights into this subtle phenomenon areoffered by additional results, presented graphically in [1].

The effect of successive device failures that are correlatedhas also been considered [3]. The existing theoretical modelscan be enhanced to also consider the correlation effectaccording to the methodology detailed in [3]. The resultsobtained demonstrate that correlated device failures have anegative impact on reliability.

44 ERCIM NEWS 120 January 2020


45ERCIM NEWS 120 January 2020

Browse, visualise, Analyse

Eu Procurement data

by Ahmet Soylu and Till C. Lech (SINTEF)

Annually, around 14% of the EU’s GDP is spent on the

procurement of goods and services by over 250,000

public authorities. Improving the effectiveness, efficiency,

transparency and accountability of government

procurement is therefore in the public’s interest. The

increasing amount of open procurement data enables us

to analyse public spending to deliver better quality and

more economical public services, prevent fraud and

corruption, and build healthy and sustainable economies.

TheyBuyForYou [L1] is a three year project, funded byEuropean Union's Horizon 2020 program, that aims to build atechnology platform consisting of a set of modular web-basedservices and APIs, to publish, curate, integrate, analyse, andvisualise an open, comprehensive, cross-border and cross-lin-gual procurement knowledge graph, including public spendingand corporate data from multiple sources across the EU.

Many of the tools in use by governments are not optimisedfor government use or are subject to restrictive contracts thatcreate unnecessary complications when it comes to pub-lishing open data. Other contracts, such as contracts fortender advertising portals, are hampering the progress oftransparency because the portals are claiming copyright overall data published in the portals, even though their public-sector clients are the authors and the data on tender opportu-nities are required to be published openly by law. The tech-nical landscape for managing such contracts is very hetero-geneous: for example, even in medium-sized cities, contractsare handled using different tools and formats across depart-ments, including relational databases, Excel spreadsheets,and Lotus Notes. This makes it difficult to achieve a high-level overview of processes and decisions. There are variousinitiatives, such as Open Contracting Data Standard (OCDS),that aim to create de-jure and de-facto standards for elec-

tronic procurement. However, these are mostly oriented toachieve interoperability (i.e., addressing communicationbetween systems), document oriented (i.e., the structure ofthe information is commonly provided by the content of thedocuments that are exchanged), and provide no standardisedpractices to refer to third parties, companies participating inthe process, or even the main object of contracts. In short,there is enormous heterogeneity in systems and processes.The Semantic Web approach has the potential to benefit theprocurement domain by allowing the reuse of existingvocabularies, ontologies, and standards.

The TheyBuyForYou project explores how procurementknowledge graphs, paired with data management, analyticsand interaction design could be used to reform four key pro-curement areas [1]: (i) economic development: facilitatingbetter economic outcomes from public spending for SMEs;(ii) demand management: spotting trends in public spendingto achieve long-term goals such as savings; (iii) competitivemarkets: promoting healthier competition and identifyingcollusions and other irregularities; and (iv) supplier intelli-gence: optimising supply chains. The project develops anintegrated technology platform with data, core services, openAPIs and online tools, which will be validated in differentbusiness cases.

The project is underway and we have integrated two high-quality datasets according to an ontology network [2], com-pany (i.e., legal entities) and procurement (e.g., tenders andcontracts) data, to form an interconnected knowledge graphfor public procurement. We ingest data from two mainproviders: OpenCorporates [L2] for supplier data (i.e., com-pany) and OpenOpps [L3] for procurement data. OpenOppshas gathered over 2,000,000 tender documents from morethan 300 publishers through web scrapping and by using openAPIs, while OpenCorporates currently has 1,400,000 entitiescollected from national registers. The data ingestion processcomprises several steps using data APIs of both providers,including data curation (e.g., handling missing values andduplicates), matching suppliers appearing in tender dataagainst company data (i.e., reconciliation), and translatingdatasets into the underlying graph data representation (i.e.,

RDF) with respect to the ontology network[3]. The current release of the knowledgegraph includes 2,300,000 triples (i.e.,records) [L4]. An example query and itsresults are depicted in Figure 1. The examplequery lists the top ten companies that consti-tute the major suppliers, in the Norwegianjurisdiction, where the jurisdiction datacomes from OpenCorporates and contractdata comes from OpenOpps. In addition tothe knowledge graph and data ingestioncomponents, we are developing a set ofonline toolkits including cross-languagedocument comparison and analytics compo-nents; anomaly detection components; acomprehensive set of guidelines for datavisualisation and interaction design; and adesign for a story-telling tool.

For the remainder of the project, we willingest more data and focus on data qualityFigure�1:�An�example�query�executed�on�the�TBFY�knowledge�graph�integrating�two�datasets.



issues, which turns out to be considerable challenge.Currently, we are working on various approaches to improvedata quality, ranging from machine learning to crowd-sourcing. Finally, we are consolidating an integrated platformwith data access APIs and online tools supporting differentdata analytics tasks. A series of business cases are being builton top of the knowledge graph and online tools to realise theaforementioned four key innovation areas.

The project consortium is composed of SINTEF (coordinator,Norway), University of Southampton (UK), OpenOpps (UK),OpenCorporates (UK), Cerved (Italy), Ministrstvo za JavnoUpravo (Slovenia), Ayuntamiento de Zaragoza (Spain), JozefStefan Institute (Slovenia), Oesía Networks SL (Spain), andUniversidad Politecnica de Madrid (Spain).

Links:

[L1] https://theybuyforyou.eu/[L2] https://opencorporates.com/[L3] https://openopps.com/[L4] http://data.tbfy.eu

References:

[1] Simperl et al.: “Towards a Knowledge Graph BasedPlatform for Public Procurement”, in Proc. of MTSR2018.

[2] Soylu et al.: “Towards an Ontology for Public Procure-ment Based on the Open Contracting Data Standard”, inProc. of I3E 2019.

[3] Soylu et al.: “An Overview of the TBFY KnowledgeGraph for Public Procurement”, in Proc. of ISWCSatellites 2019.

Please contact:

Ahmet Soylu, SINTEF, [email protected]

More on 5G: Millimetre-

Waves

by Gregor Dürrenberger (FSM) and Harry Rudin

The tumult surrounding 5G continues in Switzerland. The

telecommunications providers are intensely promoting

5G; industry is waiting for 5G, citizens are concerned

about potential health issues associated with 5G, and the

press is full of articles about 5G. In a few years, the

technology will also use millimetre-wave transmission. In

anticipation of this in June this year the Swiss Research

Foundation for Electricity and Mobile Communication

(FSM) at the Swiss Federal Institute of Technology in

Zurich organised a millimetre-wave workshop.

The millimetre-wave workshop, held in Zurich in June,2019, touched on many topics, most of which were directlyor indirectly related to the interaction between millimetre-waves and humans, and the ways that 5G might impact us.Many of the charts from the presentations are available at[L1].

Millimetre-waves have been around us for a long time: Theyhave many commercial applications such as material surfacetreatment and physical measurements and are also used inradar for contactless detection systems, such as speed detec-tors and weather radar. Other applications are directlyrelating to people, including medical applications, such ascomputer tomography scans. An overview of commerciallyavailable applications was given by Christoph Baer, Ruhr-University in Bochum, and of biomedical applications byMaxim Zhadobov, French National Centre for ScientificResearch, Rennes.

The public’s main concern is the effect of 5G millimetre-waves on our health and wellbeing. We know that electro-magnetic fields interact with the human body. The type ofinteraction depends on frequency, the magnitude of effectson field-strength or power. In the microwave and millimetre-wave domains exploited by mobile communication tech-nologies, the relevant interaction mechanism is believed tobe energy absorption.

It is known, as Maria Rosaria Scarfi from the Bio-EM Lab inNaples pointed out, that overexposure may have detrimentalhealth effects on the tissue level like coagulations, structuralprotein damage or disruption of organic functions. However,regulatory frameworks limit human exposure and protectagainst such risks. In terms of non-thermal biological effectsthat might occur at low exposure levels, no experimentallyverified mechanism is known for either microwaves or mil-limetre-waves.

Scarfi presented a literature analysis of published in-vitroexperiments with millimetre-waves. Most articles reportedeffects, many of them beneficial in nature. However, giventhe low quality of most studies, especially regardingdosimetry, it is premature to draw conclusions. This makes italmost impossible to separate thermal from potential non-thermal effects. Furthermore, no consistent correlationbetween effects and exposure conditions (frequency, powerdensity, exposure duration) could be identified. This mightreflect the real situation, however, it is possibly an artefactstemming from experimental choices or publication bias.Scarfi concludes “The available studies do not provide suffi-cient and adequate information for meaningful safety assess-ments” and she sees an urgent need for further, more care-fully controlled studies.

Figure�1:�Attentive�Audience�at�ETH�during�FSM�Millimeter-Wave

Workshop.�Photo:�FSM.



The complexity and challenges of bioresearch, includingmodelling, with millimetre-waves was further elaborated byJoachim Oberhammer from Karolinska Institute, Stockhom,Daniel Erni, University of Duisburg-Essen (presentationgiven by Sonja Huclova) and Sven Kühn from IT’IS, Zurich.The talks focused on the skin, which is the relevant interac-tion tissue for wavelengths in the millimetre range. Huclovashowed that we need multi-scale skin models in order toproperly assess energy absorption of microscopic structuressuch as hair channels and sweat ducts, and more empiricalresearch into the tissue parameters for sub-THz frequencies.This call for more accurate numerical models and measure-ment data was corroborated by Kühn’s presentation, whichstressed limitations in the current exposure guidelines thatrequire review and updating.

Oberhammer’s presentation introduced millimetre-waves asdiagnostic tools, especially for skin-cancer screening.Currently, no technology is available to perform screeningsthat are efficient in terms of both time and money. Thesensor-concept presented is a solution that has already beentested. Its rationale: because melanoma have higher watercontent compared to healthy tissue or benign tumours, scan-ning absorption rates of electromagnetic millimetre-waveenergy with probes allows identification of malignanttumours very quickly, even by non-dermatologists.Oberhammer described the first micromachined, high-reso-lution probe with a size below 1 mm and an operating fre-quency of 100 GHz.

In an earlier paper, “5G: A View from Switzerland” [L2] theauthors wrote about the public's concerns regarding 5G fre-quency allocation and non-ionizing radiation limits as theypertain to mobile wireless communication. Three presenta-tions at the workshop were devoted to this topic. PhilippeHorisberger from the Swiss Federal Office ofCommunications (BAKOM) reported on the many institu-tions working on the standardisation of 5G, including fre-quency band allocation on the global level. The organisationsworking on these tasks include, among many others, theIEEE and ITU at the international level, ETCI and CEPT inEurope, and many industry and national standardisationorganisations. The roadmap for harmonisation of the 26 GHzband in Europe was presented as well as the planned alloca-tion of a total of a bandwidth of 33 GHz in the domainbetween 24 and 86 GHz scheduled for the World RadioConference in November 2019.

Specifically related to health issues is the ICNIRP, theInternational Commission on Non-Ionizing RadiationProtection. Martin Röösli, member of the commission, dis-

Figure�2:�Sub-THz�frequency�allocation�plan�for�mobile�services.�Source:�workshop-presentation�of�Philippe�Horisberger,�BAKOM.�

cussed the current (preliminary) status of the revision ofICNIRP's guidelines in the millimetre-microwave region.The guidelines are specifically aimed at protecting humansfrom potentially detrimental health risks under realistic cir-cumstances for two specific populations: the general publicand those people particularly exposed through their profes-sional work. The presentation focused on both microwaveand millimetre-wave frequencies. Concerning the latter,ICNIRP intends to make the following major adjustments(proposal status): first, maximum energy absorption by thebody (technically defined as Specific Absorption Rate SAR)must be respected up to 300 GHz (formerly: 10 GHz).Second, the regulation on maximum power density startswith 6 GHz (formerly 10 GHz). Finally, SAR should extenddown to 6 GHz (formerly 10 GHz), so, between 6 and 10GHz both SAR and power density limits must be respected.

We look forward to the imminent revision of the ICNIRPguidelines.

IEEE recommends similar, but not identical, standards asICNIRP. Work is underway to consolidate the recommenda-tions of those two bodies. Ralf Bodemann, EMF consultantfrom Munich and former Chairman of IEEE ICES aka IEEEStandards Coordinating Committee 39 (SCC 39), explainedthe current status of the revision process: only marginal revi-sions to the current guidelines are under discussion, and theyinclude the addition of exposure standards for consumergoods.

Workshop participants received a research update on mil-limetre-waves, from technical to biological issues, to thestatus of ongoing discussions on regulatory revisions. Animportant outcome of the 1.5 day gathering at ETH in Zurichwas that the scientific community needs additional tissuedata and more detailed models to properly simulate the inter-action between millimetre-waves and the skin, and it needsmore high-quality in-vitro bio-experiments in order to verifyor to falsify the as yet unclear status of the many effectsreported in literature.

Links:

[L1] https://kwz.me/hK5[L2] https://ercim-news.ercim.eu/en117/special/5g-a-view-

from-switzerland

Please contact:

Gregor Dürrenberger, Swiss Research Foundation forElectricity and Mobile Communication, ETH Zurich,+41 44 632 [email protected]



Announcements

ANNOUNCEMENT and CALL for PAPERS

FMICS 2020: 25th International

Conference on Formal Methods

for Industrial Critical Systems

Austria, 2-3 September 2020

FMICS is the yearly conference organised by the homony-mous ERCIM working group. The aim of the conferenceseries is to provide a forum for researchers who are inter-ested in the development and application of formal methodsin industry. FMICS brings together scientists and engineersactive in the area of formal methods and interested inexchanging their experiences in the industrial usage of thesemethods. The conference series also strives to promoteresearch and development for the improvement of formalmethods and tools for industrial applications. This 25th edi-tion will be celebrated in a special way.

QONFEST 2020FMICS 2020 is part of the QONFEST umbrella conferencecomprising also CONCUR, FORMATS, and QEST, alongwith workshops and tutorials, from August 31 to September5, 2020. The topics covered by QONFEST are theory, formalmodelling, verification, performance evaluation and engi-neering of concurrent, timed, industrial critical, and othersystems. FMICS 2020 features a special track on "FormalMethods for Security in IoT".

Submission and PublicationThe paper submission deadline is May 15, 2020. Acceptedpapers will be included in the Conference Proceedings pub-lished in Springer's Lecture Notes in Computer Science inthe subline on Formal Methods. Authors of selected paperswill be invited to submit an extended version of their paperto a special issue of the "International Journal on SoftwareTools for Technology Transfer".

Conference Chairs: Maurice ter Beek, ISTI-CNR (PC co-chair), Dejan Ničković,AIT (PC co-chair), Ezio Bartocci, TU Wien (General chair).

Invited Speakers

• Thomas Henzinger, IST (joint keynote with CONCURand QEST)

• Stefan Resch, Thales• Roderick Bloem, TU Graz (joint keynote with CONCUR

and FORMATS)

More information:

ERCIM Membership

ERCIM membership is open to research institutions(including universities). By joining ERCIM, your researchinstitution or university can participate in ERCIM’s activi-ties and contribute to the ERCIM members’ commonobjectives playing a leading role in Information andCommunication Technology in Europe.

“Through a long history of successful research

collaborations in projects and working groups and

a highly-selective mobility programme, ERCIM has

managed to become the premier network of ICT

research institutions in Europe. ERCIM has a

consistent presence in EU funded research

programmes conducting and promoting high-end

research with European and global impact. It has a

strong position in advising at the research policy

level and contributes significantly to the shaping of

EC framework programmes. ERCIM provides a

unique pool of research resources within Europe

fostering both the career development of young

researchers and the synergies among established

groups. Membership is a privilege.”Dimitris Plexousakis,

Director of ICS-FORTH, Greece

About ERCIMERCIM – the European Research Consortium forInformatics and Mathematics – aims to foster collaborativework within the European research community and toincrease cooperation with European industry. Founded in1989, ERCIM currently includes 17 leading research estab-lishments. ERCIM is able to undertake consultancy, devel-opment and educational projects on any subject related toits field of activity.

ERCIM members are centres of excellence across Europe.ERCIM is internationally recognized as a major representa-tive organization in its field. ERCIM provides access to allmajor Information Communication Technology researchgroups in Europe and has established an extensive programin the fields of science, strategy, human capital and outreach.ERCIM publishes ERCIM News, a quarterly high qualitymagazine and delivers annually the Cor Baayen Award tooutstanding young researchers in computer science orapplied mathematics. ERCIM also hosts the Europeanbranch of the World Wide Web Consortium (W3C).

[email protected]


ERCIM “Alain Bensoussan”

Fellowship Programme

The ERCIM PhD Fellowship Programme has been estab-lished as one of the premier activities of ERCIM. The pro-gramme is open to young researchers from all over theworld. It focuses on a broad range of fields in ComputerScience and Applied Mathematics.

The fellowship scheme also helps young scientists toimprove their knowledge of European research structuresand networks and to gain more insight into the working con-ditions of leading European research institutions.

The fellowships are of 12 months duration (with a possibleextension), spent in one of the ERCIM member institutes.Fellows can apply for second year in a different institute.

Why to apply for an ERCIM Fellowship?The Fellowship Programme enables bright young scientistsfrom all over the world to work on a challenging problem asfellows of leading European research centers. In addition, anERCIM fellowship helps widen and intensify the network ofpersonal relations and understanding among scientists. Theprogramme offers the opportunity to ERCIM fellows:• to work with internationally recognized experts;• to improve their knowledge about European research

structures and networks;• to become familiarized with working conditions in leading

European research centres;• to promote cross-fertilization and cooperation, through the

fellowships, between research groups working in similarareas in different laboratories.

ConditionsCandidates must: • have obtained a PhD degree during the last eight years

(prior to the year of the application deadline) or be in thelast year of the thesis work with an outstanding academicrecord;

• be fluent in English; • have completed their PhD before starting the grant;• submit the following required documents: cv, list of publi-

cations, two scientific papers in English, contact details oftwo referees.

The fellows are appointed either by a stipend (an agreementfor a research training programme) or a working contract.The type of contract and the monthly allowance/salarydepends on the hosting institute.

Application deadlinesDeadlines for applications are currently 30 April and 30September each year.

Since its inception in 1991, over 500 fellows have passedthrough the programme. In 2019, 53 young scientists com-menced an ERCIM PhD fellowship and 79 fellows havebeen hosted during the year. Since 2005, the FellowshipProgramme is named in honour of Alain Bensoussan, formerpresident of Inria, one of the three ERCIM founding insti-tutes.

http://fellowship.ercim.eu

Having a ERCIM Postdoc was agreat opportunity! I moved toanother country, I met new peopleand I worked in machine learningapplied to industry. This definitelychanged my career. Thanks a lot!

Rita Duarte PIMENTELFormer ERCIM Fellow

ERCIM fellowship certainly contribute toenhance current and future careerprospects and opportunities. In addition,ERCIM fellowship gives chances forincreasing the network of contacts andsetting up international collaborations,mainly through short visiting periods inother ERCIM institutes and conferences. Istrongly recommend this Fellowship, whichis also a great life experience.

Ahmed ELARABYFormer ERCIM Fellow

It was a holistic experience, and therealization of the experience andknowledge I gained only grows each time Ilook back on it. ERCIM fellowship is aunique opportunity that needs to begrasped, that shapes your future and cangive you experiences that will inspire youfor many years.

Naveed Anwar BHATTIFormer ERCIM Fellow

Call for Proposals

dagstuhl Seminars

and Perspectives

Workshops

Schloss Dagstuhl – Leibniz-Zentrum für

Informatik is accepting proposals for

scientific seminars/workshops in all

areas of computer science, in particu-

lar also in connection with other fields.

If accepted the event will be hosted inthe seclusion of Dagstuhl’s well known,own, dedicated facilities in Wadern onthe western fringe of Germany.Moreover, the Dagstuhl office willassume most of the organisational/administrative work, and the Dagstuhlscientific staff will support the organ-izers in preparing, running, and docu-menting the event. Thanks to subsidiesthe costs are very low for participants.

Dagstuhl events are typically proposedby a group of three to four outstandingresearchers of different affiliations. Thisorganizer team should represent a rangeof research communities and reflectDagstuhl’s international orientation.More information, in particular, detailsabout event form and setup as well as theproposal form and the proposing processcan be found on

https://www.dagstuhl.de/dsproposal

Schloss Dagstuhl – Leibniz-Zentrum fürInformatik is funded by the German fed-eral and state government. It pursues amission of furthering world classresearch in computer science by facili-tating communication and interactionbetween researchers.

Important Dates• Proposal submission:

April 1 to April 15, 2020• Notification: July 2020• Seminar dates: Between March 2021

and February 2022 (tentative).

Events


HORIZON 2020

Project Management

A European project can be a richlyrewarding tool for pushing yourresearch or innovation activities to thestate-of-the-art and beyond. ThroughERCIM, our member institutes haveparticipated in more than 90 projectsfunded by the European Commission inthe ICT domain, by carrying out jointresearch activities while the ERCIMOffice successfully manages the com-plexity of the project administration,finances and outreach.

Horizon2020:Howcanyougetinvolved?The ERCIM Office has recognized ex-pertise in a full range of services,in-cluding:• Identification of funding opportunities• Recruitment of project partners

(within ERCIM and throughournetworks)

• Proposal writing and projectnegotiation

• Contractual and consortiummanagement

• Communications and systems support• Organization of attractive events,

from team meetings to large-scaleworkshops and conferences

• Support for the dissemination ofresults.

How does it work in practice? Contact the ERCIM Office to presentyour project idea and a panelof expertswithin the ERCIM Science Task Groupwill review youridea and provide recom-mendations. Based on this feedback,theERCIM Office will decide whether tocommit to help producing yourproposal.Note that having at least one ERCIMmember involvedis mandatory for theERCIM Office to engage in a project.Ifthe ERCIM Office expresses its interestto participate, it willassist the projectconsortium as described above, either asprojectcoordinator or project partner.

Please contact:

Peter Kunz, ERCIM [email protected]

Tim Berners-Lee

Launches Global

Action Plan to Prevent

"digital dystopia"

Tim Berners-Lee, together with leadersfrom government, business, civic groupsand citizens from across the world, haslaunched the world’s first-ever globalaction plan to halt the increasing misuseof the web and ensure it is protected as aforce for good. The inventor of the weburged widespread backing of the plan, theContract for the Web, which sets out con-crete actions governments, companiesand individual citizens can take to ensurea web that is safe, empowering and foreveryone. The failure of global actors todefend the free and open web, he warned,risks a “digital dystopia” of entrenchedinequality and abuse of rights.

The web has proved one of the mostpowerful tools we have to change livesfor the better. However, its benefit tohumanity is at risk due to a growing dig-ital divide in access to the web and anincreasing number of online threats,including election interference, onlineharassment, threats to privacy and thespread of disinformation. Tim Berners-Lee challenged governments and compa-nies to show leadership in addressing thethreats facing the web: “The Contract forthe Web gives us a roadmap to build abetter web. But it will not happen unlesswe all commit to the challenge.Governments need to strengthen lawsand regulations for the digital age.Companies must do more to ensure pur-suit of profit is not at the expense ofhuman rights and democracy. And citi-zens must hold those in power account-able, demand their digital rights berespected and help foster healthy conver-sation online. It’s up to all of us to fightfor the web we want.”

The Contract for the Web, led byBerners-Lee’s World Wide WebFoundation, has been backed by over150 organisations,. Thousands of indi-viduals, hundreds of organisations andthe governments of Germany, Franceand Ghana signed up to the Contract’sfounding principles.

Individuals and organisations are

asked to endorse the Contract at:

https://contractfortheweb.org/

david Chaum and Guido van

Rossum awarded dijkstra

Fellowship

CWI has awarded David Chaum and Guido van Rossum forthe first time the newly established honorary title ‘DijkstraFellow’. Chaum ‘forefather of blockchain’ is well known forhis groundbreaking research in the area of Privacy andCryptology and the development of digital currency. VanRossum created the world-famous programming languagePython almost 30 years ago, when he was working at CWI.

With the Dijkstra Fellowship CWI aims to recognize theimpact of their groundbreaking work on mathematics andcomputer science. The Fellowship is also a recognition forthe contribution they made to CWI’s reputation as a pioneerin the area of Mathematics and Computer Science. The hon-orary title is named after former CWI researcher Edsger W.Dijkstra, who was one of the most influential scientists in thehistory of CWI by - next to other accomplishments - devel-oping the shortest path algorithm.

Guido van Rossum (left) is awarded with the DijkstraFellowship for the impact of his programming languagePython. Python is currently one of the most used program-ming languages in the world. Python is highly used for pro-gramming applications in the field of AI. The language isavailable open source and has a globally active user com-munity.

David Chaum (right), mathematician and computer scientistreceives the honorary title “Dijkstra Fellow” for hisadvanced research in the area of Privacy and Cryptology. Bysetting up the Cryptology research group at CWI in the early80’s, he brought modern cryptology to Europe. He is interna-tionally recognized as the founder of digital currencies andblockchain technology. Spin-off Digicash emerged from hisresearch work. Chaum developed eCash, the precursor of thebitcoin, under the flag of Digicash.

https://www.cwi.nl/news/blogs/interview-david-chaum-201cblockchain-will-decentralize-power201d

https://www.cwi.nl/news/blogs/interview-guido-van-rossum-201cid-rather-write-code-than-papers.201d


W3Cx Introduction to Web

Accessibility – New Online Course

The W3C and UNESCO Institute for InformationTechnologies in Education (IITE) have launched a newW3Cx course: Introduction to Web Accessibility. Designedfor technical and non-technical audiences, including devel-opers, designers, content authors, project managers, peoplewith disabilities, and others, this course helps get a strongfoundation in digital accessibility to make websites and appswork well for people with disabilities, meet internationalstandards, and provide a better user experience for everyone.The course starts on 28 January 2020 and is self-paced. Readmore in the course description and enroll soon!

This course is based on the open curricula of the W3C WebAccessibility Initiative (WAI), a framework for developingcourses using role-based modules that build on and extendthe Web Accessibility Initiative (WAI) accessibility trainingresources. These curricula are available to anyone to use asan authoritative foundation for creating courses, and arebeing developed through the consensus process of the W3CEducation and Outreach Working Group (EOWG) with sup-port from the European Commission (EC) funded WAI-Guide Project (Grant 822245).

https://www.edx.org/course/web-accessibility-introductionhttps://www.w3.org/WAI/curricula/

HAL is Opening up to Software

As a result of the collaboration between Software Heritage,HAL-Inria and the CCSD (Centre for Direct ScientificCommunication), the open archive HAL is opening up to anew type of scientific data: software. Henceforth, researcherstherefore have the possibility of depositing source codewhilst contributing to the software heritage built up bySoftware Heritage. HAL is an open archive where authorscan deposit scholarly documents from all academic fields.

https://hal.archives-ouvertes.fr/https://www.softwareheritage.org/

Phot

o: C

WI/

Inge

Hoo

glan

d

ERCIM is the European Host of the World Wide Web Consortium.

Institut National de Recherche en Informatique et en AutomatiqueB.P. 105, F-78153 Le Chesnay, Francewww.inria.fr

VTT Technical Research Centre of Finland LtdPO Box 1000FIN-02044 VTT, Finlandwww.vttresearch.com

SBA Research gGmbHFloragasse 7, 1040 Wien, Austriawww.sba-research.org/

Norwegian University of Science and Technology Faculty of Information Technology, Mathematics and Electri-cal Engineering, N 7491 Trondheim, Norwayhttp://www.ntnu.no/

Universty of WarsawFaculty of Mathematics, Informatics and MechanicsBanacha 2, 02-097 Warsaw, Polandwww.mimuw.edu.pl/

Consiglio Nazionale delle RicercheArea della Ricerca CNR di PisaVia G. Moruzzi 1, 56124 Pisa, Italywww.iit.cnr.it

Centrum Wiskunde & InformaticaScience Park 123, NL-1098 XG Amsterdam, The Netherlandswww.cwi.nl

Foundation for Research and Technology – HellasInstitute of Computer ScienceP.O. Box 1385, GR-71110 Heraklion, Crete, Greecewww.ics.forth.grFORTH

Fonds National de la Recherche6, rue Antoine de Saint-Exupéry, B.P. 1777L-1017 Luxembourg-Kirchbergwww.fnr.lu

Fraunhofer ICT GroupAnna-Louisa-Karsch-Str. 210178 Berlin, Germanywww.iuk.fraunhofer.de

RISE SICSBox 1263, SE-164 29 Kista, Swedenhttp://www.sics.se/

Magyar Tudományos AkadémiaSzámítástechnikai és Automatizálási Kutató IntézetP.O. Box 63, H-1518 Budapest, Hungarywww.sztaki.hu/

TNOPO Box 968292509 JE DEN HAAGwww.tno.nl

University of CyprusP.O. Box 205371678 Nicosia, Cypruswww.cs.ucy.ac.cy/

ERCIM – the European Research Consortium for Informatics and Mathematics is an organisa-

tion dedicated to the advancement of European research and development in information tech-

nology and applied mathematics. Its member institutions aim to foster collaborative work with-

in the European research community and to increase co-operation with European industry.

INESCc/o INESC Porto, Campus da FEUP, Rua Dr. Roberto Frias, nº 378,4200-465 Porto, Portugal www.inesc.pt

I.S.I. – Industrial Systems InstitutePatras Science Park buildingPlatani, Patras, Greece, GR-26504 www.isi.gr

SIMULAPO Box 1341325 Lysaker, Norwaywww.simula.no

Figure�1:


Subscribe to ERCIM News and order back copies at https://ercim-news.ercim.eu/

special theme: educational technology · educational technologies, “edtech”, illustrating the...

Documents