slow intelligence systems session and panel

45
1 Slow Intelligence Systems Session and Panel

Upload: kellsie

Post on 16-Jan-2016

49 views

Category:

Documents


0 download

DESCRIPTION

Slow Intelligence Systems Session and Panel. Panelists. Erland Jungert Francesco Colace Tiansi Dong Shi-Kuo Chang (Moderator). Outline. Motivation Introduction to SIS Application: Ontological Filters Application: Topic/Trend Detection Discussion. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Slow Intelligence Systems Session and Panel

1

Slow Intelligence Systems

Session and Panel

Page 2: Slow Intelligence Systems Session and Panel

2

Panelists

• Erland Jungert• Francesco Colace• Tiansi Dong• Shi-Kuo Chang (Moderator)

Page 3: Slow Intelligence Systems Session and Panel

3

Outline

• Motivation• Introduction to SIS• Application: Ontological Filters• Application: Topic/Trend Detection• Discussion

Page 4: Slow Intelligence Systems Session and Panel

4

Motivation: Common Characteristics ofNew Generation Information Systems

• Connected• Multiple sourced• Knowledge-based• Personalized• Hybrid

Page 5: Slow Intelligence Systems Session and Panel

5

Smarter Planet• We are all now connected - economically,

technically and socially. Our planet is becoming smarter via integration of information scattered in many different data sources: from the sensors, on the web, in our personal devices, in documents and in databases, or hidden within application programs. Often we need to get information from several of these sources to complete a task. Examples include healthcare, science, the business world and our personal lives. (Quoted from Josephine M. Cheng, IBM Fellow and Vice President of IBM Research)

Page 6: Slow Intelligence Systems Session and Panel

6

(courtesy of IBM)

Page 7: Slow Intelligence Systems Session and Panel

7

Hybrid Intelligence• While processor speed and storage capacity

have grown remarkably, the geometric growth in user communities, online computer usage, and the availability of data is in some ways is even more remarkable. Hybrid Intelligence offers great opportunities we have to harness this data availability to build systems of immense potential. While today s large scale systems are evolutionarily based on the distributed computing technologies envisioned in the 70 s and 80 s, sheer scaling has led to many unanticipated challenges. (quoted from Alfred Z. Spector, Vice President, Research and Special Initiatives, Google, USA)

Page 8: Slow Intelligence Systems Session and Panel

8

Hybrid IntelligenceUsers and computers doing more than either could

individually (quoted from Alfred Z. Spector, Google).

Page 9: Slow Intelligence Systems Session and Panel

9

Slow Intelligence Systems• Slow Intelligence Systems are general-

purpose systems characterized by being able to improve performance over time.

• A slow intelligence system is a system that (i) solves

problems by trying different solutions, (ii) is context-

aware to adapt to different situations and to propagate

knowledge, and (iii) may not perform well in the

short run but continuously learns to improve its

performance over time.

Page 10: Slow Intelligence Systems Session and Panel

10

Slow Intelligence Systems• Slow Intelligence Systems are general-

purpose systems characterized by being able to improve performance over time

through a process involving • Enumeration

Page 11: Slow Intelligence Systems Session and Panel

11

Slow Intelligence Systems• Slow Intelligence Systems are general-

purpose systems characterized by being able to improve performance over time

through a process involving • Enumeration• Propagation

Page 12: Slow Intelligence Systems Session and Panel

12

Slow Intelligence Systems• Slow Intelligence Systems are general-

purpose systems characterized by being able to improve performance over time

through a process involving • Enumeration• Propagation• Adaptation

Page 13: Slow Intelligence Systems Session and Panel

13

Slow Intelligence Systems• Slow Intelligence Systems are general-

purpose systems characterized by being able to improve performance over time

through a process involving • Enumeration• Propagation• Adaptation• Elimination

Page 14: Slow Intelligence Systems Session and Panel

14

Slow Intelligence Systems• Slow Intelligence Systems are general-

purpose systems characterized by being able to improve performance over time

through a process involving • Enumeration• Propagation• Adaptation• Elimination

• Concentration

Page 15: Slow Intelligence Systems Session and Panel

15

Slow Intelligence Systems• Slow Intelligence Systems are general-

purpose systems characterized by being able to improve performance over time

through a process involving • Enumeration• Propagation• Adaptation• Elimination

• Concentration• Slow Decision Cycle to complement Fast

Decision Cycle

Page 16: Slow Intelligence Systems Session and Panel

16

Slow Intelligence Systems

• A SIS continuously learns, searches for new solutions and propagates and

shares its experience with other peers.

• From the structural point of view, a SIS is a system with multiple decision

cycles such that actions of slow decision cycle(s) may override actions of quick decision cycle(s), resulting in

poorer performance in the short run but better performance in the long-

run.

Page 17: Slow Intelligence Systems Session and Panel

17

SIS Basic Building Block (BBB)

Page 18: Slow Intelligence Systems Session and Panel

18

Advanced Building Block (ABB)

Page 19: Slow Intelligence Systems Session and Panel

19

SIS is a component-based systembuilt from BBBs and ABBs

Page 20: Slow Intelligence Systems Session and Panel

20

The SIS Testbed for Healthcare Systems

Page 21: Slow Intelligence Systems Session and Panel

21

Ontological Filters for Slow Intelligence Systems

Shi-Kuo Chang, Emilio Zegarra, Francesco Colace and Massimo De Santo

Page 22: Slow Intelligence Systems Session and Panel

22

Production of personalized or custom-tailored goods or services to meet consumers' diverse and changing needs

SIS Application to Product Configuration

Page 23: Slow Intelligence Systems Session and Panel

23

Ontological Filter and Slow Intelligence System

Figure 6 - Ontological Filter and the Slow Intelligent System

Page 24: Slow Intelligence Systems Session and Panel

24

A Scenario

• A customer would like to buy a Personal Computer in order to play videogames and surf on the internet.

• He knows that he needs an operating system, a web browser and an antivirus package.

• In particular, the user prefers a Microsoft Windows operating system. He lives in the United States and prefers to have a desktop. He also prefers low cost components.

Page 25: Slow Intelligence Systems Session and Panel

25

Ontological Transform for Product Configurator

Page 26: Slow Intelligence Systems Session and Panel

26

Building Topic/Trend Detection System based on Slow Intelligence

Chia-Chun Shih & Ting-Chun PengInstitute for Information Industry

Taipei, Taiwan

Page 27: Slow Intelligence Systems Session and Panel

27

• An online trend detection system requires careful resource allocation and automatic algorithm adaptation to process huge size of heterogeneous data.

• This research adopts Slow Intelligence, which provides a framework for systems with insufficient computing resources to gradually adapt to environments, to response the challenges.

• Four Slow Intelligence subsystems are proposed, and each subsystem targets a challenge in designing online topic/trend detection systems.

Page 28: Slow Intelligence Systems Session and Panel

28

Introduction • Topic Detection and Tracking (TDT)

– Initiated by DARPA at 1996– discover the topical structure in

unsegmented streams of news reporting as it appears across multiple media

– Tasks:• Topic Detection• Topic Tracking• First Story Detection• Story Segmentation• Link Detection

Page 29: Slow Intelligence Systems Session and Panel

29

Topic/Trend Detection System

• Objective– Detect current hot topics and to predict future hot

topics based on data collected from Social Media

• Three components– Crawler & Extractor: Collect data and extract

information from Social Media– Topic Extractor: Detect hot topics from a set of text

documents– Trend Detector: Detect trends (future hot topics)

based on currently available data

Crawler & Extractor

Topic Extractor

Trend Detector

SocialMedia

Current Hot topics

Future Hot topics

Page 30: Slow Intelligence Systems Session and Panel

30

Topic/Trend Detection System

• Crawler & Extractor

(cont’d)

Web dataDB

WebCrawler

HTMLdocuments

InformationExtractor

* Extract articles and metadata (title, author, content, etc) from semi-structured web content

User’sKeywords of

Interests

Topic Extractor

Social Media

Textdocuments

Crawler & Extractor

Page 31: Slow Intelligence Systems Session and Panel

31

Topic/Trend Detection System

• Topic Extractor

(cont’d)

Web dataDB

Topic WordExtraction

Topic WordClustering

Hot topicextraction

Currenttopics

CurrentHot topics

Topic Extractor

• Apply TF-IDF scheme to generate Top-N topic words for each document

• Apply clustering algorithm to cluster topic words into topic groups. The topic groups are treated as “topics” • Apply aging theory to

find hot topics

Page 32: Slow Intelligence Systems Session and Panel

32

Topic/Trend Detection System

• Trend Detector

(cont’d)

Trend Detector

Currenttopics

Trend EstimationAlgorithms

Topic Trend(Future Hot Topics)

Page 33: Slow Intelligence Systems Session and Panel

33

T/TD System with Slow Intelligence

• Four complexities of designing online topic/trend detection systems

• 1. It is unlikely to collect all web data based on limited amount of computing

.resources The system needs to develop data collection strategies which can concentrate limited resources on collecting important web data.

Page 34: Slow Intelligence Systems Session and Panel

34

T/TD System with Slow Intelligence

• 2. Many computation methods are available for estimating trends. If parameter settings are also taken into account, there are too many combinations to choose. Furthermore, Internet is a changing environment, which means current best solution may not perform well in the future. The system needs to automatically (or at least quasi-automatically) find best solution from many alternatives in a changing environment.

(cont’d)

Page 35: Slow Intelligence Systems Session and Panel

35

T/TD System with Slow Intelligence

• 3. The crawler needs to revisit websites to collect up-to-date data in hourly or daily intervals. Each site has different amount of to-be-update data and different policy to restrict frequent access, which are unknown beforehand. The system needs to find feasible data collection schedule based on past experience.

(cont’d)

Page 36: Slow Intelligence Systems Session and Panel

36

T/TD System with Slow Intelligence

• 4. Any changes in web pages may disrupt Extractors. It needs automatic repair mechanism for Extractors if many websites are being monitored. The repair mechanism needs to detect errors of Extractors, find alternatives, and choose the best solution from alternatives to fix the disrupted Extractors.

(cont’d)

Crawler &

Extractor

Page 37: Slow Intelligence Systems Session and Panel

37

T/TD System with Slow Intelligence

1. SIS to help restrict the range of data collection

(cont’d)

Knowledge of data

Knowledge of algorithm

Page 38: Slow Intelligence Systems Session and Panel

38

T/TD System with Slow Intelligence

2. SIS to help select and adapt trend detection algorithms

(cont’d)

Page 39: Slow Intelligence Systems Session and Panel

39

T/TD System with Slow Intelligence

3. SIS to help scheduling Crawler

(cont’d)

Page 40: Slow Intelligence Systems Session and Panel

40

T/TD System with Slow Intelligence

4. SIS to help adapt Extractors

(cont’d)

Page 41: Slow Intelligence Systems Session and Panel

41

Enumerator Adaptor Eliminator Concentrator

Slow Intelligence System Building Blocks

Crawler & Extractor Topic Extractror Trend Detector

Topic/Trend Detection System

SIS system for scheduling Crawlers

SIS system for Selecting Trend Estimation MethodSIS System for

Focused Crawling

SIS system for adapting extractors

Enumerator Adaptor Eliminator Concentrator

Slow Intelligence System Building Blocks

Crawler & Extractor Topic Extractror Trend Detector

Topic/Trend Detection System

SIS system for scheduling Crawlers

SIS system for Selecting Trend Estimation MethodSIS System for

Focused Crawling

SIS system for adapting extractors

Page 42: Slow Intelligence Systems Session and Panel

42

Discussion• There are a large number of

intelligent systems, quasi-intelligent systems and semi-intelligent systems that are "slow". Distributed intelligence systems, multiple agents systems and emergency management systems are mostly slow intelligence systems that exhibit the characteristics of multiple decision cycles.

Page 43: Slow Intelligence Systems Session and Panel

43

Discussion (continued)

• Since time is relative, "slow" intelligence systems for some can also be "fast" for others.

• A slow intelligence system can evolve

into a fast intelligence system.

• A framework for knowledge-based software engineering.

Page 44: Slow Intelligence Systems Session and Panel

Q&A

Page 45: Slow Intelligence Systems Session and Panel

The End