the role of computational intelligence in developing countries: …shura.shu.ac.uk/15780/1/mwitondi...
TRANSCRIPT
The role of computational intelligence in developing countries: data access, manipulation and interpretation - natural drivers for socio-economic transformation
MWITONDI, Kassim
Available from Sheffield Hallam University Research Archive (SHURA) at:
http://shura.shu.ac.uk/15780/
This document is the author deposited version. You are advised to consult the publisher's version if you wish to cite from it.
Published version
MWITONDI, Kassim (2017). The role of computational intelligence in developing countries: data access, manipulation and interpretation - natural drivers for socio-economic transformation. In: Computational Intelligence for Societal Development in Developing Countries (CISDIDC), Sheffield Hallam University, 17 February 2017. (Unpublished)
Copyright and re-use policy
See http://shura.shu.ac.uk/information.html
Sheffield Hallam University Research Archivehttp://shura.shu.ac.uk
1
The Role of Computational Intelligence in Developing Countries
Data Access, Manipulation and Interpretation - Natural Drivers for Socio-Economic Transformation
Kassim S. Mwitondi (PhD)Sheffield Hallam University; Computing and Communication Research Centre
[email protected] [email protected]://www.shu.ac.uk/about-us/our-people/staff-profiles/kassim-mwitondi
2
Presentation outline/Sign-posting PART I: WHAT this seminaris about– Tackling Challenges
Semantics: Computational Intelligence and Development
Underlying Challenges and Objectives
PART II: WHY? Motivation, justification, current state
Challenges and opportunities
Transforming societies via data, information and knowledge
PART III: HOW dowe do it?
Frameworks for accessing and sharing data, tools and skills
Interdisciplinary approach; knowledge transfer partnerships
POTENTIAL OUTCOMES: Data sharing frameworksinhealth, education, agriculture, tourism, international trade,revenue collection, infrastructure, governance & other sectors.
Concluding remarks, Discussions, Questions and Answers
3
Logical and Lexical Semantics of the Theme The meaning, presupposition,implication and the
relationshipamongthewordsof thetheme… Unlike mathematicsor statistics, Computational
Intelligence is notanestablishedscience
It still has no establishedfoundations as it is aportfolio of tricks and methods (formal andinformal) usedto addresschallengesmankindfaces.
Typically, the methodsare "probabilistic" seekingto deal with uncertaintyin the sameway as humanreasoning - making use of inexact/incompletedata/knowledgein adaptivecontrolenvironment.
4
There has never been a clear distinction between"developing" and "developed" countries. The InternationalMonetaryFund(IMF) distinguishesbetween"advanced" and"emerging" marketeconomieswithoutanystrict criteria.
Despite labelling 159 countriesas DCs, the United Nations(UN) doesn’t haveanofficial definitionof whataDC is.
The World Trade Organisation(WTO) lets membersdefinethemselvesas"developing"or "developed."
The World Bank previouslyuseda cut-off point the bottomtwo-thirds of gross national income (GNI) asDCsandfromlast year,it hasremovedfrom its DevelopmentIndicatorsthedistinctionbetweenthetwo concepts.
We can seewhy. Qatar, Kuwait and Singapore all haveahigherpercapitaincomethantheUS.
Logical and Lexical Semantics of the Theme
5
Computational Intelligence and Developing Countries DeanandKanazawa(1989) provideanaturallink of CI to DC in thelatter's
currentcontextof socio-economic,political andcultural transformation. Inparticular,the authorsdescribea modelof causalreasoningthat accountsfor knowledgeconcerningcause-and-effectrelationships.
Their motivation was basedon the premisesthat "…reasoning aboutchange requires predicting how long a proposition, having become true,will continue to be so." Consequently, in the absenceof perfectknowledge,it is naturalto expectanagentto be"…constrained to believethat a proposition persists indefinitely simply because there is no way forthe agent to infer a contravening proposition with certainty."
Their foundationsof causalreasoningspanbackto the ancientGreeksandforth into recentadvancesin computingas it addresses classical problemsin temporal reasoning - which fits nicely in thedomainof DC.
In linking CI to DC, we assertthat "intelligence controls reasoning” anddecisionmaking-it is thereforeavital attributeweneedto "develop."
Our challenges opportunities, strengths and weaknesses lie within thedomains of what we know, what we don’t, what we can and can’t do.
Other things being equal, we succeed because “we know” and we failbecause we lack knowledge and so Alexander the Great and the ancientGreeks had to consult the Oracle at Delphi before going to war.
We still consult fortune tellers, consultants and spiritual leaders beforegetting married, before making landmark decisions.
Bioinformatics and computational biology have opened new avenues inbiomedical Sciences – they gave us the human genome.
Business Intelligence – we believe, is a way we should do business today. Criminology utilises new technologies in controlling our borders (not
walls) – we can track and model emotions; lie detection etc. Food and water are in short supply – space/terrestrial weather can help. World populations are growing. We track and model migration trends Satellite maps of disappearing ice and extreme weather events warn us
of the impact of climate change – unless you believe it is a Chinese hoax.
6
CORRELATION BETWEEN KNOWLEDGE AND DEVELOPMENT
An easy to understandGCSE poverty cycle
Knowledge is key toaddressing each node
Each node consists ofmeasurable data attributes
These must be harnessed,stored, modelled & shared
DCs’ approach to Big Datamust be adaptive
Computational intelligenceto identify and address realchallenges on the ground
DCs need reliable datarepositories, shared acrossdisciplines and boundaries
Galileo Galilei: Measurewhat is measurable andmake what is notmeasurable measurable
7
DC’s CORE CHALLENGE: TACKLING THE POVERTY CYCLE
Source: https://whsgcsegeographysupport.wordpress.com/the-development-gap/
8
MULTIVARIATE NATURE OF DEVELOPMENT Eachof the 17 goalsis
multivariate and theyare highly correlated
In a typical DC, thesedata attributes are high-dimensional, multi-faceted and fragmented
We cancluster… We canclassify… We canassociate… We canpredict… Or canwe? Do weknow… WHO is doing or did WHAT? WHERE? WHEN? HOW?
Source: http://www.africa.undp.org/content/rba/en/home/sustainable-development-goals.html
9
The Global OpenData for Agriculture& Nutrition initiative(GODAN, 2017)asserts that thesolution to thechallenge of hungerand malnutrition“Zero Hunger lieswithin existing, butoften unavailable,agriculture andnutritiondata”.
Hence, it promotesOpen Data policytomake informationabout agriculture andnutrition availableinorder to help addressthe challenge ofglobalfood security.
Source: http://www.africa.undp.org/content/rba/en/home/sustainable-development-goals.html
ADDRESSING GOAL #2 (ZERO HUNGER) THROUGH OPEN DATA
10
Optimal linking of farmingcommunities with potentialmarkets/customers
Visualisation of spatio-temporal aspects of localfarming, changing patterns
Reducing wastageby linkingwith families, schools,Maasais and Sukumas…
Its scheme, Open GeospatialScience is built upon thepremises that“…scientificknowledge…develops morerapidly and productivelyifopenly shared…”
Its key ingredients areenshrinedin the Principles–“Open Source geospatialsoftware, Open data, Openstandards ,Open educationalresources, and Open accessto research publications.
Source: http://www.africa.undp.org/content/rba/en/home/sustainable-development-goals.html
ADDRESSING GOAL #2 (ZERO HUNGER) THROUGH OPEN DATA
11
Source: http://www.africa.undp.org/content/rba/en/home/sustainable-development-goals.html
UNESCO-UIS (2016) reports anumberof challengesin learningandassessment.
Most DCs’ educationalsystemseitherhavecolonial influencesorhave been adopted from“developing” countries. Put thesetogether and cross-nationallycomparable data becomesdifficult/impossibleto attain.
Prado and Marzal (2013)highlight the importanceof datain societyby recommendingthatlearning institutions at all levelsenhancetheir contributionto data& informationliteracy.
Sets of core competenciesandcontents that can serve as aframework of reference forinclusion in data and/orinformationliteracyprograms.
ADDRESSING GOAL #4 (QUALITY EDUCATION) THROUGH OPEN DATA
12
Hence, the General Message and Questions are… Need interdisciplinary
learning, sharing data,tools and skills.
Do what we can donow, with what wehave–data, skills, tools
Some governmentscall upon people topray for rain duringdry seasons.We do…
Can we record/sharedata on rainfall?
Can we use themtooptimise farming?
Provide clean water? Can we teach basic
association rulesto theyoungsters based onthese phenomena?
YES,WE CAN…Source: http://www.africa.undp.org/content/rba/en/home/sustainable-development-goals.html
13
Source: http://www.africa.undp.org/content/rba/en/home/sustainable-development-goals.html
Big Data?The digital divide stilldefines the global imbalance(Mwitondi, 2009).
DCs must fully utilise existingdataflow infrastructurealongsidemodernisationprogrammes.
Forming adaptive researchsynergiesbetweenthe academiaandtheprivate(KTP, morelater).
What have the DCs got? TheJapanesehave very little in theform of land and resourcesbutplentyin creativity& innovation.
Industrialisation derives fromSTI, it transforms agriculture,transport,utility supply, markets(home& abroad),mind setsand,above all, it promotes brandnames– asdid Galaxy/iPhone…
ADDRESSING GOAL #9 (INDUSTRY, INNOVATION & INFRASTRUCTURE)
14
Source: http://www.africa.undp.org/content/rba/en/home/sustainable-development-goals.html
Reforms in legislation,technology, infrastructure andmind sets– crucial humanfactorscome in the form of “education”and“good governance”.
These associations must beuncoveredandusedas inputs intodevelopmentstrategies.
If overlooked,the industrialisationequationbecomesunsolvable.
Everythingthat we do in life – asindividuals, families, ministriesorcountries we set off from thesimplestto thecomplex.
Institutional arrangementamongvariousbodiesis inevitable.
Sharing what is known, what isaccessible helps consolidatewhateverstrategiestheremaybe.
ADDRESSING GOAL #9 (INDUSTRY, INNOVATION & INFRASTRUCTURE)
15
WHAT IS AVAILABLE AND ACCESSIBLE?High-speed fibre optic connectivity has provided robust Internetbackbone and hence general ICT development across the world.
The project ideanever took offdue tofinancial, politicaland technical hurdles
The hurdles couldprobably have beenovercome, had therebeen “useful” datasources and potentialflows across thecontinentat the time.
Africa One: It wasperceivedas a greatideaat the turn of thelast millennium.
16
WHAT HAVE THE DCs ACHIEVED? AND WHAT ELSE CAN BE DONE?
Twenty years on, a numberof submarine and terrestrial fibre optic cablesconnect African countriesto the Internet and more are planned for major cities!
They are expectedto revolutionise ICT trends around the continent and enhanceits socio-economic & cultural transformation.In major cities?
Beneficiaries from these developments have mainly been telecoms companies–users benefitting via telephone banking, e-Health and educational schemes.
Could and shouldwedomore with respectto the development goals? Surely.
Source: https://mybroadband.co.za/news/broadband/106353-here-is-who-controls-the-internet-in-south-africa.html
17
Many Challenges and Opportunities Still Lie Ahead Many parts of the DCs are yetto significantly address existing challenges and potential
opportunities.In Africa, there are many who believe that potential benefits are constrainedby its spatial coverage andsosatellite technologyis the best wayto reach rural areas.
Unfortunately, both options stillaren’t standardin most DCs– including Kenya - acountry viewedasa leaderin African technology innovation.
Source: http://www.bbc.co.uk/news/business-32079649
Isizwe considers satellite andmicrowave technologiesas thefuture, with Wi-Fi, 3G and 4Gproviding extensionsto smallbusinesses and homes.
The Isizwe Project in SouthAfrica focuses on wirelesstechnologies to bring freeinternet to everyone in SouthAfrica by installing wi-fihotspotsin low-income areas.
This assertion is probablysupportedby the huge success ofmobile banking across manyparts of the developing world.
Key prerequisites are policiesGVT commitment), data, skills,tools, design and management.
ChenandDahlman(2004) assessedtheeffectsof knowledgeoneconomicgrowth of 92 countriesover forty years(1960-2000)and concluded that human capital, domestic innovation,technologicaladaptation,andthe level ICT infrastructurehadasignificanteffectson long-termeconomicgrowth.
Globally, ICT R&D is gearedtowardsproductivity, innovationandprofitability – hence,proprietaryin nature. In the DCs ICTeducation produces “systems supporters” who are oftendetachedfrom mainstreamsocio-economicphenomena.
This trend createsgaps betweenthe elite & the rest of thepopulation; between the academia & industry. It formsknowledgeclusterswithin thesedomainsandso restrictsaccessto data,knowledge,toolsandskills. Hence,aswe talk of CI andDC, couplingandcohesionof theseclustersarefundamental.
18
PUTTING KNOWLEDGE AND DEVELOPMENT INTO PERSPECTIVE
With all thehypeaboutBig DataandICT, the digital divide stilldefinessocio-economicimbalancesbetweenthetwo hemispheres(Mwitondi, 2009). Bridging this gap is an uphill and multi-dimensionaltaskthatrequiresco-ordinatedinterventions.
GOVERNMENTSMUST TAKE THE LEAD Weathington (2017) underlines the need for CEOsin companiesto lead
Big Data initiativesin organisations– aspartof corporate strategy. DC Governments must play key rolesin promoting open sourceaspartof
their“development strategy” otherwise they are doomedto fail. A top-down approachis recommended. The political elites– at the very
top, must champion these initiatives.It seemsto beworking in Rwanda. KNOWLEDGE TRANSFERPARTNERSHIPS
To bridge this gap, DCs must optimally utilise existing data flow potentials– novel and conventional and the best wayto achieve thisis by
…forming adaptive research synergies between the academia, privatesector and other public institutions. These initiatives need funding.
Usebusiness-political synergiesto enforce contractual clauses…19
JUST HOW CAN THE GAPS BE NARROWED?
20
How? Collectively Set Vision, Values And Measurable Objectives (VVMO)
Intersecting socio-economic, cultural andenvironmental developmentis centralto development.
Central & local governments, NGOs and individualsmust come together with shared VVMO. These mayvary but they mustbeadapted along the way
How we "shape" our development spectrumisextremely centralto attaining VVMO.
Our VVMO require invariance– this is a challenge. Geometry describes "shape"as what remains after
removing the effectsof "translation and rotation" - thatis, clearing perturbations, crisesor disagreements.
Coon (1946) describes universalityof human societiesfrom various perspectives. Developmentis “invariant”anddata onattributes that describeit are crucial.
21
Source: http://visualizenomalaria.org
SOME OF THE STEPS ALREADY TAKEN
The project“boosts economies” bycutting malaria treatment costs(currently overUS$12bna year)
It frees up scarce public healthresources (40% is spentonmalaria)
It strengthens education (50% ofabsenteeismdueto malaria).
Its visualisation tools are loaded… Most DCs run similar projects -
funded differently and quite oftennot talking to eachother.
Data and/or information gainedoften get lost between projects,ministries, funding arrangements.
Bill and Melinda Gates Foundation,the WB, USAID, DFID, JICA etc
We should retain datato helpdescribe our universality withrespectto “development”).
22
SOME OF THE STEPS ALREADY TAKEN
Fancy namesto attractadmissions? Maybe...
What are their outcomes? Mostof startatUni-level How wide-spread? How accessible? Do they helpuscurb real
life challenges? Yes… Coulddomore? Yes… Whatis wrong?
The universalityof global phenomena stipulates thatourstrategies for tackling real-world challenges start early
http://www.bbc.co.uk/news/education-38938519
23
STEPS THAT NEED TO BE TAKEN Interdisciplinary approachto problem solving
must be adopted from much earlier inprofessional careers than the current practiceis.
Merging disparate disciplines basedof theiranalytical requirements, educational systemscanbuild interdisciplinary bridges across curricula.
Unifying conceptsin various disciplinesis not novel to research communities.Mwitondi (2003) discusses the linkbetween statistics & computer science.From similar premiseswe can definecategories of data literacy,competencies and resources needed.
24
1. High Poverty in areas surrounding National Parks
2. Limited local benefits3. Poor infra/high travel costs4. Increase in concentration of
economic activity in the ecosystem:
(a) Agricultural expansion(b) Pressures on carrying
capacity (c) Deforestation
5. Poaching6. Illegal logging
Examples of Composite Challenges in Southern Tanzania
About 56% of flow to the Mteraand Kidatupower plants – make 50% of total TZ hydropower capacity
Rapidly Expanding Irrigation in and around the UsanguPlains (BRN)
Reduced Environmental Flows, affects Mteraand Kidatu
Tourism significance – RuahaNP is the core of the southern circuit but without water no wildlife/tourism
Downstream economic activities Poverty impacts; levels of water use/stress puts livelihoods at risk
Human-wildlife conflicts expanding during, especially in dry seasons
Settlements without schools25
Further Examples of Composite Challenges in Southern Tanzania
27
Measuring What is Measurable
A Jamaican-born ladynamed Sue Stolbergerhasbeen regularly measuringthe depth of the GrandRuaha & compiling variousfact files for over 12 years.
Sue is not a researchscientist. Sheis inspired byher love for art and nature.
Yet, she maintains thrillingrudimentary data repository.
How about empoweringlocal schoolsto carry outthis exercise and share data?
On-going initiatives aresupportedby the central andlocal GVTs,WB and others
28
SUPPORTING OPEN SOURCEDC GOVERNMENTS MUST PLAY KEY ROLES
Proprietary tools have thrivedfor a reason– R&D investment. Open software developmentis proneto more risks than proprietary– open
testing exposesit to hackers; its vulnerabilityis compoundedby the commonphenomenonin programming - component dependency.
Open source vulnerability managementis and will remain a key concern.With a variety of contributors from around the world, developmentmanagement teams may find themselves pushedto the edgeof propriety.
NASA code (https://code.nasa.gov) and R (https://www.r-project.org/about.html) (http://developer.r-project.org/)
have stringent measuresin place and they are well-funded. Weathington (2017) underlines the need for CEOsin companiesto leadBig
Data initiativesin organisations– aspartof corporate strategy. DC Governments must play key rolesin promoting open sourceas part of
their“development strategy” otherwise they are doomedto fail. A top-down approachis recommended. The political elites– at the very top,
must champion these initiatives.At leastit seemsto beworking in Rwanda.
29
SUPPORTING OPEN SOURCE APPLICATIONSADAPTIVE EDUCATIONAL & VOCATIONAL PROGRAMMES NEEDED
R&D funding in mostDCsis predominantly foreign-dependent. NASA Europa Challenge– OpenCitySmart– for developing
tools for city-related infrastructure managementis based onEurope's INSPIRE Directiveto guide project development.
While data management tools maybe universal, traffic controlin Dar, Lagos or Mumbai may need innovative approaches…
…innovations that don’t necessarily derive existingsophisticated traffic control systems yousee in metropolitanLondon, Tokyoor New York, but, possibly portable systems ofsmart apps focusingon “what works”. Examples already exist
The same approach should be adoptedin tacklingeachof the 17goals. But remember, they are inseparable andsothey require acoherent, unified approach– requiring yourCI skills and tools.
30
Typical Examples of Collaborative Setting of VMOs
Enabling comparative studies as a way of positively contributing towards understanding the clinical and/or pathogenetic features of diseases like
Sickle Cell Anaemia (SCA)Systemic Lupus ErythematosusVarious types of cancerLymphatic FilariasisMalariaHIV/AIDS
Collect/manage data for modelling disease patterns.Investigation of race-specific diseases.Developing a tracking index for various diseases that could be used as a benchmark for intervention programmes.
PR
ED
ICT
IVE
MO
DE
LL
ING
DE
TE
CT
ING
NA
TU
RA
L
STR
UC
TU
RE
S IN
DA
TA
COMPARATIVE STUDIES
SHARING DATA, RESEARCH FINDINGS, SKILLS & TOOLS
31
SUPPORTING OPEN SOURCE APPLICATIONSADAPTIVE EDUCATIONAL & VOCATIONAL PROGRAMMES NEEDED
R&D funding in mostDCsis foreign-funded and often owned. NASA Europa Challenge– OpenCitySmart– you can join the
initiative. But note thatit is inspired by Europe's INSPIREDirective andsoit maynot bedirectly transferable.
Data management tools may be universal, but traffic controlinDar, Lagos or New Delhi may need adaptive solutions.
You should not seekto emulate or improve upon thesophisticated traffic control systemsyou seein London, Tokyoor New York, but focus on“what works”, while you remain“smart”. These are likelyto be portable systems– like themobile apps, currentlyin usein major African cities.
Adopt the same philosophyin tackling eachof the 17 goals.They are correlated, hence require coherent, unified approach.
Dynamicsof the Grand Ruaha (km 475 long over 83,970squarekilometres)canbe tracked and monitored without the need forsatellites. You just needto beproactive, innovative and“open”.
32
BEARING IN MIND THE FOREGOING CONSIDERATIONS, YOU CAN THEN APPLY YOUR SKILLS AND ALL THE NOVEL TOOLS YOU HAVE DEVELOPED
33
BEARING IN MIND THE FOREGOING CONSIDERATIONS, YOU CAN THEN APPLY YOUR SKILLS AND ALL THE NOVEL TOOLS YOU HAVE DEVELOPED
34
References and Bibliography1) Chen, D. and Dahlman, J. (2004). Knowledge and Development: A Cross-Section Approach.
Policy Research Working Paper;No.3366. World Bank, Washington, D.C.2) Coon, C. (1946). The Universalityof Natural Groupingsin Human Societies; The Journalof
Educational Sociology, Vol.20, No. 3,pp. 163-168, American Sociological Association.3) Dean, T. and Kanazawa, K. (1989). A model for reasoning about persistence and causation
Computational Intelligence; Vol 5, Issue 2,pp 142–150.4) Dreher, A., Gaston, N. and Martens, P. (2008). Measuring Globalisation: Gauging Its
Consequences; Springer, ISBN-13: 978-0387740676.5) GODAN (2017). Global OpenDataInitiative; http://www.godan.info/about6) Kuznetsov, Y. (2006). Diaspora Networks and the International Migrationof Skills: How
Countries Can Drawon Their Talent Abroad;WB Publications, ISBN-13: 978-08213664797) Mwitondi, K. S. (2009); Tracking the Potential, Development, and Impactof Information and
Communication Technologiesin Sub-Saharan Africa; Abook chapter in: Science,Technology, and Innovation for Socio-economic Development: Success Stories from Africa;International Council for Science (ICSU-ROA), ISBN978-0-620-45741-5.
8) Prado, J. and Marzal, M. (2013). Incorporating Data Literacy into Information LiteracyPrograms: Core Competencies and Contents; Vol.63, Issue 2,pp. 123–134, ISSN0024-2667
9) UNESCO-UIS (2016). Laying The FoundationTo Measure Sustainable Development Goal 4;Sustainable DevelopmentDataDigest, ISBN978-92-9189-197-9.
10) Weathington, J. (2017). Why CEOs Must Lead Big Data Initiatives;http://www.techrepublic.com/article/why-ceos-must-lead-big-data-initiatives/