TMP 419 D
TMP 419 D
TMP 419 D
a
e-mail: j.h.johnson@open.ac.uk
216 The European Physical Journal Special Topics
1 Introduction
FuturICT is a visionary ten year programme to deliver new science and technology
to explore, understand and manage our complex and connected world. This paper
explores how FuturICT can help to revolutionise education in the decade ahead.
FuturICT aims for an ICT-enabled quantum shift in human knowledge capital-
ising on the current data revolution, new methods and models to use those data in
large distritbuted simulators, and new forms of individual and social social behaviour
enabled by an evolving internet incorporating new intelligent technologies and more
natural human-computer interaction. FuturICT seeks to harness the disruptive power
of such technology with a vision for evolving knowledge in the social sciences and a
vision for the emerging science of complex systems to create an ethically grounded
platform for decision making and policy:
“This system will be able to act as a ‘Policy Simulator’ or ‘Policy Wind Tunnel’,
allowing people to test multiple options in a complex and uncertain world, and
produce pluralistic perspectives of possible outcomes. The framework would
analyse data on a massive scale and leverage them with scientific knowledge,
thereby giving politicians and decision-makers a better understanding to base
their decisions on. Through the concept of a socially inclusive Participatory
Platform, FuturICT will extend such capabilities to empower citizen, commu-
nities, small businesses, and NGOs, creating a whole ecosystem of new appli-
cations and forms of social and economic participation. In the long run, this
would enable every one of us to explore the possible or likely consequences
of even barely imaginable scenarios, effectively helping us to see just a little
around the corner into possible futures.” [1]
We propose that FuturICT could catalyse a much needed revolution educational
provision. Moreover, it will itself be a major beneficiary of that revolution: to achieve
its goals FuturICT will need new generations of highly educated and well-trained
people across the disciplines and technologies. As a ‘man-on-the-moon’ federated
Big Science project FuturICT will build on several hundred teams of scientists
worldwide and involve many thousands of people. Many of these will be in their
teens when the programme begins. They will need training and education in subjects
currently at research frontiers. They will need interdisciplinary education providing
solid foundations within and across disciplines. They will need the ability to self-
educate as new ideas and knowledge emerge at unprecedented rates. They will need
the ability to communicate their scientific discoveries rapidly and effectively to fel-
low scientists and those who will use the new science in the private and public sectors.
“In the field of education and training the mission of the European
Commission is to reinforce and promote lifelong learning.” . . . Education and
training policy has gained particular momentum with the adoption of the Eu-
rope 2020 strategy, the EU’s overarching programme focusing on growth and
jobs. ... Recognising that lifelong learning is key to both jobs and growth and
the participation of everyone in society, EU Member States and the European
Commission have strengthened their political cooperation through the strate-
gic framework “Education and Training 2020”. . . . “Since 2007, the European
Commission has integrated its various educational and training initiatives un-
der a single umbrella, the Lifelong Learning Programme. The objective of the
programme is to enable individuals at all stages of their lives to pursue stim-
ulating learning opportunities across Europe.” [2]
At an international level [3] “The mission of the UNESCO Education Sector is to:
Despite these aspirations, millions of people have inadequate access to high quality
education or the outcome of their educational is disappointing, as we now discuss.
The problem of limited access. The OECD [4] classifies levels of education as
pre-primary (minimum entry 3 years), primary (entry age 6–7 years, 6 years dura-
tion), lower secondary (following primary education for 3 years), upper secondary
(students have usually completed 9 years of education and are generally 15–16 years
old), post-secondary non-tertiary (duration 6 months to 2 years), and tertiary subdi-
vided to tertiary-type A (3 years including university ‘first’ degrees and some masters),
tertiary-type B (2 years with a focus on practical , technical or occupational skills),
and Advanced degree programmes (3 year advanced research programmes, including
Ph.D.s).
In most OECD countries education is compulsory at primary and secondary levels
so that “virtually everyone in the OECD area has access to at least 13 years of formal
education” ([4] page 293). Furthermore “it is estimated that an average of 59% of
today’s young adults will enter tertiary-type A (largely theory-based programmes)
and 19% will enter tertiary-type B (shorter, largely vocational) programmes over
their lifetimes” ([4] page 308), and “an estimated 2.6% of today’s young adults will
enter advanced research programmes” ([4] page 309).
In contrast to this high level of access to education in the most developed coun-
tries, it remains an aspiration for many others. “An increasing number of countries aim
for universal participation in secondary education. The social returns on investment
are greater than in higher education regardless of the income level of the country.” ...
“While participation at the secondary level has grown significantly in many countries,
equitable access and completion - as well as the quality and relevance of secondary ed-
ucation - represent major challenges” [5]. Table 1 shows the enrolment rates in primary
education for various regions of the world with sub-Saharan Africa below 60% in 2009.
218 The European Physical Journal Special Topics
Table 1. Enrolment rates in secondary education (source: UNESCO [5], page 10).
Change in primary Enrolment rate
Region population 1999–2009 1999 2009
Arab States 17.3% 77% 86%
Central and Eastern Europe −21.0% 94% 94%
Central Asia −19.9% 94% 93%
East Asia and the Pacific 14.7% 94% 94%
Latin America and the Caribean −3.1% 93% 95%
North America & Western Europe −2.5% 97% 96%
South and West Africa 28.2% 79% 91%
Sub-Saharan Africa 59.2% 59% 77%
World 8.6% 84% 90%
This relatively low level of participation in primary education has obvious implications
for higher levels of education:
“The [UNESCO Global Education Digest] shows that broader access to sec-
ondary education, however, represents a serious challenge in many parts of the
world. The gross enrolment ratio (GER) in lower secondary education increased
from 72% to 80% worldwide between 1999 and 2009, with notable increases
in the Arab States and sub-Saharan Africa. Yet despite this progress, the par-
ticipation rate for this level of education remains very low in sub-Saharan
Africa at 43%. In addition, one-third of the world’s children still live in coun-
tries where lower secondary education is formally considered compulsory but
where the commitment is not met. This is especially the case in South and
West Asia. More equitable access to secondary education is another important
challenge. Between 1999 and 2009, the GER for girls increased from 69% to
79% in lower secondary and from 43% to 55% in upper secondary education
worldwide. However, the Arab States and sub-Saharan Africa still faced se-
rious gender disparities at the lower secondary level, while disparities at the
upper secondary level intensified in South and West Asia and sub-Saharan
Africa. The Digest also examines patterns of educational attainment, out-of-
school young adolescents, classroom environments, teachers and financing of
secondary education.” ([5] page 3).
“The failure of our educational system to meet the needs of our nation’s most
disadvantaged children is disturbing. Despite Britain’s international reputation
as a home of educational excellence and our economy’s global significance, our
nation has one of the highest levels of educational inequality in the Western
world. The Educational Failure Working Group has examined why huge in-
vestment in education has failed to reverse declining social mobility and the
persistent underachievement of disadvantaged children. ... we can no longer
tolerate the underachievement and frustrated potential of disadvantaged chil-
dren.” [6]
Participatory Science and Computing for Our Complex World 219
Thus two of the richest countries in one of the richest regions of the world are aware
of deficiencies in their educational systems. A report from the European Commission
shows that this applies to Europe in general: “One-quarter of young people under the
age of 15 only attain the lowest level of proficiency in reading; 15% of young people
aged 18-24 leave school prematurely; only 78% of 22-year-olds have completed their
upper secondary education; the level of interest in some subjects, such as science and
mathematics, is low.”1
Impoverished global education data sets
As noted above, many countries pay significant attention to the comparisons of educa-
tional achievement provided by the OECD through the Programme for International
Student Assessment (PISA). PISA is a considerable achievement, providing an ev-
idence base to inform domestic educational policy. There is, however, a significant
critique among educational researchers regarding the quality of its methodology, and
the extent to which its data can support some of the conclusions and policy initiatives
that PISA’s league tables provoke in some countries [7]. Moreover, while PISA en-
ables a degree of international comparison through the use of widely available proxy
indicators of learning, it continues to reveal deep intractable challenges in education
(such as embedded disadvantage linked to geography, economics and ethnicity), but
lacks the depth and resolution needed to provide an understanding of the mechanisms
driving the patterns it surfaces.
There is a pressing need to assemble and curate an internationally comparable
data set which can better inform our understanding of those systemic relationships
shown by educational improvement researchers to be critical. These include:
Table 2. Annual expenditure per student by educational institutions for all services, by
level of education (2008). (Source: OECD[4] page 209).
Primary Secondary Tertiary
$14,000 $20,000 – Luxembourg
$11,000 $13,000 $19,000 Norway
$11,000 $ 9,000 $10,000 Iceland
$10,000 $11,000 $18,000 Denmark
$10,000 $12,000 $30,000 United States
$10,000 $12,000 $15,000 Austria
$ 9,000 $10,000 $20,000 Sweden
$ 9,000 $18,000 $22,000 Switzerland
$ 9,000 $ 9,000 $15,000 United Kingdom
$ 9,000 $ 9,000 $ 9,000 Italy
$ 9,000 $10,000 $15,000 Belgium
$ 9,000 $11,000 $16,000 Ireland
$ 8,000 $ 9,000 $15,000 Japan
$ 7,000 $11,000 $17,000 Netherlands
$ 7,000 $10,000 $14,000 Spain
$ 7,000 $ 9,000 $16,000 Finland
$ 7,000 $ 9,000 $15,000 Australia
$ 7000 $ 9000 $14000 OECD average
$ 6,000 $10,000 $14,000 France
$ 6,000 $ 9,000 $16,000 Germany
$ 6,000 $ 7,000 $10,000 New Zealand
$ 6,000 $ 7,000 – Estonia
$ 6,000 $ 8,000 $ 9,000 Korea
$ 6,000 $ 6,000 $13,000 Israel
$ 5,000 $ 7,000 $10,000 Poland
$ 4,000 $ 4,000 $ 7,000 Hungary
$ 4,000 $ 4,000 $ 7,000 Slovak Republic
$ 4,000 $ 6,000 $ 8,000 Czech Republic
$ 3,000 $ 3,000 $ 8,000 Chile
$ 2,000 $ 4,000 $ 7,000 Argentina
$ 2,000 $ 2,000 $ 4,000 Mexico
$ 2,000 $ 2,000 $ 7,000 Brazil
$ 1,000 $ 1,000 – Indonesia
– $ 9,000 $ 9,000 Slovenia
– $ 8,000 $19,000 Canada
– $ 4,000 $ 7,000 Russian Federation
– – $ 4,000 China
Three features emerge from this analysis. The first is that education is not inclusive
with some countries failing to provide the most elementary levels of education for
some of their population. The second is that even in those countries that do provide
primary and secondary education for all their population the outcome can be deficient.
The third is that lack of provision of education, especially in Africa, is not due to
disproportionately low spending on education.
Thus the educational challenges faced in Europe and worldwide are immense: ed-
ucation is expensive but this does not guarantee quality, and the investment made
in mass education is ineffective for large minorities. Put simply, our traditional ap-
proaches to education are failing, and the obvious answer of allocating more resource
to education may be neither necessary nor sufficient to provide high quality education.
222 The European Physical Journal Special Topics
Fig. 1. Education spending in 2009 as a percentage of GDP. (Source UNESCO [5], page
73).
Clearly there is a need for more effective education, and this presents enor-
mous challenges but great opportunities for innovation enabled by the FuturICT
programme.
One size fits all versus personalised education
Some of the most effective teaching is done by highly competent experts giving in-
struction on a one-to-one basis with their pupils. But this approach does not scale to
large numbers of learners and much more cost-effective methods of teaching and learn-
ing need to be invented. Today mass education constrains cost using a production line
approach. The raw materials are mostly young people aged four to twenty four formed
into batches according to their age and geography. Curriculum, the blueprint, is often
created at national level by the ministry of education. Within the education factory,
be it called a school, college or university, individual students are aggregated by age
and geography into groups of twenty, thirty, forty or even hundreds to be processed
in the same way. The point is made by Sir Ken Robinson in a highly entertaining
animated lecture to the Royal Society of Arts given in 2010 [93]2 : “ The current sys-
tem was designed and conceived in a different age ... in the intellectual culture of the
enlightenment and in the economic circumstances of the industrial revolution ... if
you are interested in the model of education you don’t start from a production line
mentality ... it’s about standardisation ... I believe we have to go in the exact opposite
direction”. In contrast to the one-size-fits-all approach, FuturICT education must be
personalised to suit the individual person at any particular time.
Silo domains versus interdisciplinary knowledge
The knowledge humankind has accumulated over millennia is divided into subject
domains and is generally taught in strictly demarcated disciplines. However the mod-
ern world is highly connected and its problems do not neatly fit any particular do-
main classification. Increasingly they involve many interacting physical and social
subsystems and their behaviour depends crucially on how the subsystems interact.
Furthermore, innovation typically involves interactions across domains. Whereas deep
2
http://www.youtube.com/watch?v=zDZFcDGpL4U
Participatory Science and Computing for Our Complex World 223
domain-based knowledge will remain essential, there is an increasing need for inter-
disciplinary education.
Complex systems science has to be an integrative science. The traditional
‘silo’ domains such as physics, chemistry, biology, medicine, psychology, sociology,
economics, geography, history, and linguistics are researched in depth, usually
independently of each other. Cutting across these vertical domains are horizontal
considerations that apply to and integrate them all. These includes the general
philosophical and epistemological questions as to how to reconstruct the dynamics
of systems from data, mathematical theories for representing system dynamics,
issues of data and statistics, and the use of ICT to collect, store, process, display
and exchange data and information. With their focus on particular domains, the
conventional sciences are incapable of modelling the dynamics of complex multilevel
systems of systems of systems, but an interdisciplinary approach is an essential
requirement for a general science of complex systems. It will always be necessary
to have domain specialists drilling deep into their subjects, but the science of
complex systems requires some scientists to be able to work across the disciplines
and their different cultures. This creates a major educational challenge. At university
level most people are educated in a single domain, and in this respect most of us
know almost nothing about almost everything [17]. To create the numbers of people
qualified to realise the FuturICT vision requires a major effort at all educational levels.
The pedagogical challenge now being recognised is greater than the fragmentation
of knowledge-systems into silos or lack of engagement. Yes, educational institutions
must teach the mastery of cross- and trans-disciplinary skills and understanding, and
yes, this must be done in highly engaging ways that connect the material to students’
lives. Beyond this, however, learners need to be taught how to learn more rapidly
and skillfully in whatever context they find themselves - much of what they learn will
date rapidly, they will have many jobs, and learning is no longer what happens at
preset times in special places: our connected world is ‘always on’, and the deepest
learning occurs when connected to one’s everyday life. This is part of learning to
manage complexity. The turbulence of today’s social and economic conditions places
unprecedented pressure on people’s capacity to deal with uncertainty and adapt to
change. Learners increasingly need to find meaning in the face of ambiguity and
conflicting voices, to critically evaluate information, and use their agency to positively
shape the local and the global communities in which they are involved [18].
While nurturing these qualities and capacities with students better equipped as
citizens for the extraordinary complexity in everyday life, a key outcome expected
from education is fitness to work. Here that the evidence is also challenging. Business
and industry report consistently that graduates - from school and university sectors -
lack the transferable skills they need to perform effectively; moreover, this covers the
entire spectrum of work, not just high-tech, ‘knowledge-intensive’ work [19–21].
Grand Challenge 1
Enable people to learn orders of magnitude more effectively than they do today.
Grand Challenge 2
Enable people to learn at orders of magnitude less cost than they do today.
Grand Challenge 3
Demonstrate success by exemplary interdisciplinary education in complexity science.
The first of the grand challenges requires explicitly defined learning outcomes which
must be measurable. Generally learning outcomes involve gaining (i) knowledge and
understanding (e.g. knowing an equation), (ii) cognitive skills (e.g. deriving that
equation), (iii) key skills (e.g. applying that equation in a new context), and (iv)
practical and professional skills (e.g. using a program to compute the equation and
documenting the application according to established practices and standards). In
this context, traditional examinations and projects test what a student has learned
by these criteria. Another key measurement is learning time. Other measurements
include learner enjoyment and engagement.
The second grand challenges involves the financial and personal costs of education.
Measuring financial costs is just part of the story, since for example, learners time is
a personal cost, as is learning without enjoyment.
The term ‘order of magnitude’ in the grand challenges is interpreted as meaning
a factor between two and ten.3 It is assumed that quantum shifts in effectiveness and
costs are achievable through new understanding and applications of ICT.
3
Making education ten times more effective and ten time less expensive in ten years is a
‘man-on-the-moon’ aspiration that some think unattainable. Indeed making education twice
as effective at half the cost would be an historic achievement, with enormous global impact.
The term ‘order of magnitude’ depends on the base of the number system. For the decimal
system it means a factor of ten, while for the binary system it means a factor of two. This
justifies our interpretation of ‘order of magnitude’ as meaning a factor somewhere between
two and ten, the former being an ‘almost achievable’ minimum target and the second being
the ‘impossible goal’ that we strive to achieve.
Participatory Science and Computing for Our Complex World 225
The third grand challenge involves demonstrating that the Education Accelerator
works and that its underlying theory and practices are valid and widely applicable. It
will be a ‘wind tunnel’ for testing new theory and methods in the context of purposeful
educational programmes.
In the short term the focus will be on education in interdisciplinary complex
system science at tertiary and postgraduate levels. If the quantum shift we seek is
attainable at all, it will be easiest to demonstrate at these higher educational levels
where one can assume good learning skills, high motivation and learner cooperation.
If the grand challenges cannot be met in this benign context, it is unlikely that they
will be met in the highly political and contentious context of primary and secondary
education. Thus our strategy is to demonstrate what works at at tertiary and post-
graduate levels in the short term, taking what is proven to all levels in the medium
and longer terms.
Complexity science is a natural focus of our educational research programme. It is
interdisciplinary (work has to be done across silo discipline boundaries), multidiscipli-
nary (there may be many disciplines) and transdisciplinary (the methods of complex
system science cut across all domains at all levels providing insights into their dy-
namics and combining them into a coherent synthesis for understanding, design and
policy). In the short term, FuturICT will need accelerated educational programmes
for the thousands of scientists needing to fill in the gaps in their knowledge of its
fundamental pillars of complexity science, social science and ICT. Thus FuturICT
itself is the natural laboratory for research into the Education Accelerator.
The use of computers in education has a long and vibrant history. Our vision is that
through it various interacting components FuturICT has the potential to create new
theories and research structures that can accelerate the discovery of new methods of
pedagogy and support new educational practice.
prediction and theory: new understandings of prediction, predictive theories science and pol-
icy: new understandings of science embedded in policy global systems science: new ways of
combining knowledge from heterogeneous domains logic: new logics able to integrate multi-
level normative, technical and vernacular forms mathematics: new mathematical structures,
new mathematical models, etc. statistics: new ways of handling data, new statistical theory,
etc. simulation: new kinds of simulation, new ways of using simulations in learning, etc.
design: new methodology for the design of dynamic complex artificial multilevel systems
As in many other fields, there is now active interest in the possibility that the
concepts and tools of complexity science hold the promise of providing a new,
more rigorous language and suites of computational tools for systemic thinking
within educational research. These could enable possible futures to be mapped,
modelled, simulated, and rendered in appropriate forms to help both researchers
and practitioners to understand and, where appropriate, choose to act differently to
achieve their intended outcomes. A central claim to be investigated in this research
programme is that complexity science provides a language for transdisciplinary
learning-centred discourse between system stakeholders, serving as reference points
for modelling and, suitably communicated and embodied in tools, for educational
leaders, and learners.
– To take another example, the evidence is that efforts to manage educational sys-
tems (whether at national or institutional level) which do not take into account
complex systems dynamics, do not result in sustained school improvement: stan-
dards in schools across the developed world are plateauing, as measured by stu-
dent outcomes [23]. There is a pressing need for management and self evaluation
processes [24] which can account for such complexity in order to facilitate, value
and enhance the breadth and range of student outcomes. The evidence emerging
from these new approaches is that systemic transformation is indeed possible (e.g.
[25–28]).
Participatory Science and Computing for Our Complex World 227
– New pedagogy for e-books: Innovative ways of teaching and learning with next-
generation e-books
– Publisher-led short courses: Publishers producing commercial short courses for
leisure and professional development
– Assessment for learning: Assessment that supports the learning process through
diagnostic feedback
– Badges to accredit learning: Open framework for gaining recognition of skills and
achievements
– MOOCs: Massive Open Online Courses
– Rebirth of academic publishing: New forms of open scholarly publishing
– Seamless learning: Connecting learning across settings, technologies and activities
– Learning analytics: Data-driven analysis of learning activities and environments
– Personal inquiry learning: Learning through collaborative inquiry and active in-
vestigation
– Rhizomatic learning: Knowledge constructed by self-aware communities adapting
to environmental conditions
Synthesis
Although listed under the headings of social science, complex systems and ICT, these
breakthroughs involve a synthesis of them all. All the breakthroughs in social science
will involve the methods of complex systems and computer science, while complex
systems science is ICT-enabled, and the ICT breakthroughs relate to complex socio-
technical systems. The intertwined nature of social science, complex systems science
and ICT is exemplified by the platforms that will be created within FuturICT.
FuturICT will build new ICT systems to collect massive data sets and mine them for
useful or meaningful information, with the capacity to self-organise and adapt to the
needs of users [94]. It will be built on three new interconnected instruments: a ‘Living
Earth Simulator’ [34], a ‘Planetary Nervous System’ [35] and a ‘Global Participatory
Platform’ [36]. The details are sketched below, as summarised in [1].
228 The European Physical Journal Special Topics
A Living Earth Simulator that will enable the exploration of future scenarios at
different degrees of detail, integrating heterogeneous data and models and employing
a variety of perspectives and methods (such as sophisticated agent-based simulations,
multi-level models, and new empirical and experimental approaches). Exploration will
be supported via a ‘World of Modelling’ – an open software platform, comparable to
an app-store, to which scientists and developers can upload theoretically informed
and empirically validated modelling components that map parts of our real world.
The Living Earth Simulator will require the development of interactive, decentralised,
scalable computing infrastructures, coupled with an access to huge amounts of data,
which will become available by integrating various data sources coming from online
surveys, web and lab experiments, and from large-scale data mining.
A Planetary Nervous System that can be imagined as a global sensor network,
where ‘sensors’ include anything able to provide data in real-time about socio-
economic, environmental or technological systems (including the Internet). Such an
infrastructure will enable real-time data mining - reality mining - and the calibration
and validation of coupled models of socio-economic, technological and environmental
systems with their complex interactions. It will even be possible to extract suitable
models in a data-driven way, guided by theoretical knowledge.
A Global Participatory Platform that will promote communication, coordination,
cooperation and the social, economic and political participation of citizens beyond
what is possible through the eGovernance platforms of today. In this way, FuturICT
will create opportunities to reduce the gap between users and providers, customers
and producers etc., facilitating a participation in industrial and social value gen-
eration chains. Building on the success principles of Wikipedia and the Web 2.0,
societies will be able to harness the knowledge and creativity of multiple minds
much better than we can do today. The Global Participatory Platform will also
support the creation of Interactive Virtual Worlds. Using techniques such as serious
multi-player online games, we will be able to explore possible futures – not only for
different designs of shopping malls, airports, or city centres, but also for different
financial architectures or voting systems.
An Innovation Accelerator will augment these three platforms. It will identify in-
novations early on, distil valuable knowledge from a flood of information, find the best
experts for projects, and fuel distributed knowledge generation through modern crowd
sourcing approaches. In particular, the Innovation Accelerator will support commu-
nication and flexible coordination in large-scale projects, co-creation, and quality as-
sessment. Hence, the Innovation Accelerator will also form the basis of the innovative
management of the FuturICT flagship. Beyond this, it will fill the vision of Europe’s
Innovation Union with life and create many new business opportunities, e.g. based
on socio-inspired innovations.
Exactly what will be the outcome of the FuturICT Knowledge Accelerator cannot be
predicted. However, we believe that building on other work on ICT-enabled changes
in education will give deep insights into social organisation and the behaviour of
individuals. This will be at the level of local social structures, institutions at meso-
levels and global structures and policies at macro-levels.
ICT is enabling much better understanding of the structure of knowledge and how
networks of knowledge evolve [37]. FuturICT will accelerate this process to provide
Participatory Science and Computing for Our Complex World 229
automated analyses of the way knowledge is structured, and so provide firm founda-
tions on which to base programmes of education and training. FuturICT will create
major new platforms for accelerating knowledge on complex social systems including
simulators enabling users to explore the future and the impact of policies. These plat-
forms will be ideally suited for formal and informal education at all levels [38]. For
example, to manage the increasing amount of educational material that is now being
stored in databases, techniques such as user modelling will be needed to ensure that
appropriate materials can be delivered to the individual users [39]. The Education
Accelerator will benefit from new methods of intelligent search and data mining that
will emerge from the FuturICT research programme.
The ability to learn things often depends on familiarity with or having mastered
earlier prerequisites. If the early stages of learning are not in place, then trying to
learn can become boring, frustrating, ineffective and alienating. Currently, students
can waste a lot of time in their learning through following educational paths not
appropriate to them as an individual at a given time.
The notion of personalised learning attempts to address these issues, with the
emergence of a new class of software known as Recommender Systems (e.g. exemplified
by the ACM Recommender Systems conference: http://recsys.acm.org). Efforts to
model aspects of educational domains and learners take a number of forms, including
user models comparing the inferred cognitive model against an ideal model (intelligent
tutoring [40]); presentation layers which then tune content dynamically if progress is
deemed to be too slow (adaptive educational hypermedia [41]); and the use of data
mining for patterns that correlate user behaviour with learning outcomes (educational
data mining [42]). More intelligent software thus seeks to provide a student experience
of targeted feedback and personalised tuition [43] [44] [45] [46].
FuturICT’s particular contribution to advancing this field will most likely be in
terms of the quantity and quality of data available about learners and their broader
social contexts (via the PNS and LES platforms introduced above), and through
new forms of participatory, collective intelligence (enabled by the GPP and other
social/semantic platforms brought by this paper’s authors [47] [48].
The ‘big data’ explosion will have a big impact on education. Centrally-managed
educational records on individuals currently hold relatively sparse data, usually asso-
ciated with performance on particular tests conducted at particular times. These data
give a rough overview of a student’s knowledge and ability but compared to what will
be possible in future, when students are able to manage their own data, and bring the
richness of their educational theory to bear on their current learning, these data are
very crude. In future students will be able to continually monitor their own learning,
giving much more subtle information on which to base the tuition they require. These
new databases will be able to provide better diagnostic information in the context
of a better understanding of the way that individuals structure knowledge. This will
inform what should be taught in remedial mode and ensure that inappropriate new
things are not presented to a student before they are ready to learn them.
site navigation by making it intelligent and adaptive to the user. They needed an
autonomic system because their student numbers were rapidly growing (the number
of registered students went from 50,000 to 500,000 in 5 years) as were the number
of topics covered by their software (all topics of French secondary education). It was
thought that a solution to their problem could lie in social swarm techniques such
as those developed by ant colonies. These have been much studied by the complex
systems community as examples of complex systems.
In the real world, ants are very efficient at finding optimal paths between their
anthill and food sources. Because their functioning has been understood and modelled
as a complex system by Jean-Louis Deneubourg and Bernard Manderick in the 1980’s,
it was possible to adapt swarm intelligence systems implemented by anthill to an
educational system used by humans into what was called the man-hill optimization
paradigm [55–59].
The idea was to extract information on the behaviour of a group of students
and use it to their benefit. The emergent properties of the artificial ant system (albeit
adapted to humans) are used to find better learning paths or better learning materials
in an autonomous way.
Educational activities are divided into courses and chapters. Courses can range
from a short training course (e.g. a course on security when using heavy machinery)
to a full academic year at a school. Inside each chapter, a graph of activities is defined
that is typically composed of theory web pages or links to fundamental contents, then
exercises or links to exercises that illustrate the presented concepts, themselves leading
to a second stage where the answers of the students are corrected (and automatically
analysed for personalised remediation and global optimization of pedagogical paths
and interest).
As in artificial ants, pheromones are automatically deposited along to graph edges
depending on success or failure to validate the attempted activities, therefore reducing
the global entropy that was initially maximal, as all teaching material available in
the graph were originally unsorted. Because artificial ant systems are among the most
efficient autonomic path emergence techniques, pedagogical paths rapidly emerge that
weave an optimal learning trail among the different activities. Students weave this
path unknowingly, through a semi-transparent interface that suggest several selected
links once an activity has been validated. The links that the web interface proposes
to the students are selected among the graph, depending on the amount of positive
or negative pheromones found on the links.
In order to evaluate students’ progress (but also the difficulty of the proposed
activities), an automatic rating system was adopted based on the Chess community
ELO rating elaborated by the mathematician A. E. Elo in [61], itself based on the
Thursone Case V model [60]. The idea is to imagine that students are competing
against activities, with the result that both students and activities have an ELO
rating. Rating between students is obtained through their competition against com-
mon activities, also leading to interesting side-effects that can be used to implement
a simple real-game scheme. If the students are shown their current ELO rating as
well as the ELO rating of the activities that are suggested to them (via the man-hill
paradigm), some will chose to compete against more difficult activities, as winning
against a difficult exercise is more rewarding than winning against a simple one.
Teaching agendas or automatic remediation can also be implemented in a simple
way, by increasing personal pedagogic pheromone rates on edges leading to courses
or activities upon decision of a pedagogic actor who can be a human teacher, or an
automatic answer analyser that could detect recurrent errors on a precise topic for
a particular student. Above a certain rate of pedagogic pheromones, the course or
activity automatically pops up into the student’s personal agenda.
Participatory Science and Computing for Our Complex World 231
5.5 The virtual classroom, virtual lecture theatre and the virtual laboratory
One reason for conventional education being expensive but ineffective is that it is
conducted in geographically distributed purpose-built spaces with specialised equip-
ment. In our universities it is common to have hundreds of students assembled in a
large purpose-built theatre listening to a more or less inspiring lecturer. This social
structure precludes interaction between individual students and the lecturer, and as-
sumes that everyone in the audience is assimilating information at the same speed.
Of course the reality is that apart from listening to the lecture students are engaging
in many activities including reading their emails, chatting and even sleeping.
This method of teaching goes back hundreds of years with varying effectiveness.
Isaac Newton was required to give lectures which he often did to an empty room.
The myth that the greatest experts make the best teachers goes back a long time, as
does the myth that face-to-face lectures are always an effective method of teaching
and learning.
Since 1969 the UK Open University4 has demonstrated that well-made teaching
materials can support self-study as effectively as face-to-face lectures. In the early
days these materials included television programmes carefully planned and scripted
to communicate the most complicated ideas in accessible ways, often with excellent
graphics and animations. They also included carefully prepared pedagogic texts that
could be followed much more easily than most conventional lectures. Open University
teaching also included HEKs – Home Experiment Kits – that allowed students to
build electronic circuits, transform their kitchen into a chemistry laboratory, explore
the physics of Newton, design and control intelligent robots, and much else besides
(e.g. dissect a sheep’s brain [69]). With many hundreds of thousands of graduates,
the Open University has shown that using new technologies can make the previously
discredited model of distance education very effective.
ICT and the internet have greatly increased the possibilities for self-study in vir-
tual spaces and social structures. The potential of this is greatly enhanced when
combined with ICT-supported social structures such as social networking enabled by
the FuturICT platforms.
4
There are many Open Universities around the world, but in this paper we refer to the
original Open University founded in the UK in 1969.
Participatory Science and Computing for Our Complex World 233
One of the best ways to learn is to teach others. Although some universities use this
approach with members of one cohort of students teaching members of subsequent
cohorts, the motivation is usually to relieve senior teaching staff of some teaching
duties. There is clearly much to be learned about the pedagogic benefits of peer-to-
peer (P2P) education for students in the roles of teacher and learner [70].
The Open University’s ATELIER-D project gives evidence that some students
play the role of a ‘broker’ between disconnected networks of students. Their critical
role is that they mediate interactions between groups and therefore have the capacity
to transfer knowledge from one network to another. This point also relates to reputa-
tion within social groups, as the role of broker essentially develops through reputation
[71].
Social networking promises a great deal for P2P education. In principle social
networking can transcend geography, culture, and social level. A poor person in one
country may give instruction to a rich person in another country, and vice-versa.
What might motivate a person to help educate another? Today the motivation is
almost exclusively professional and financial - we teach because it is out job, or we
teach because we want to promote the knowledge of our social group or professional
society. This is different to students around the world supporting each other without
financial inducement.
Experience of student e-conferencing at the Open University shows that some
students like helping others to learn. In any student group it is usual for one or more
students to emerge who are willing to answer questions and direct their colleagues
to study materials that may help them. Presumably this is driven by the esteem the
individual receives within the group and other factors such as personal satisfaction
in answering questions, or a sense of social worth. It is to be expected that our
understanding of the personal motivation of individuals to help others will become
better understood, enabling a huge social resource to be harnessed for education at
all levels.
As an example, the website MathOverflow is run by research mathematicians
wanting to coordinate interest and research into open problems in mathematics:
MathOverflow’s primary goal is for users to ask and answer research level
math questions, the sorts of questions you come across when you’re writing
or reading articles or graduate level books. Of course, individual questions
don’t have to be worthy of an article, and they don’t have to be about new
mathematics. A typical example is, “Can this hypothesis in that theorem be
relaxed in this way?” [72]
A problem with P2P education is the possibility of students giving each other incorrect
guidance, promoting incorrect views and spreading misunderstanding. How can we
know that a student, or anyone else, is competent to teach another? Traditionally the
teacher is ‘qualified’ and therefore judged competent to teach. In social networking
a great deal of work has been done on reputation systems where an individual earns
a ‘reputation’ for giving good or bad advice. There are many variants on this idea,
but in its simplest form the reputation of an individual emerges from the opinions of
people who have interacted with them, e.g. if many students find a particular person
helpful the reputation of that person will improve.
Such reputation systems are already used in commercial systems, e.g. Amazon
gives many of its books and products a one-to-five star rating depending on the (freely
contributed) reviews of customers. There are of course issues such as the competence
of the reviewer, malicious reviewing and so on, but the theory and technology of
reputation systems is improving rapidly.
234 The European Physical Journal Special Topics
The conventional motivations for learning include: parent, family and social pressure;
love of learning; expectation of better job prospects or promotion in one’s career;
and peer rivalry and competition. These pressures work well for many students most
of the time, but they cause problems for some students some of the time, and they
are completely ineffective for a minority of learners for whom demotivating factors
can be stronger, e.g. discouragement or hostility to learning by parents, family, and
social groups; hatred of learning, because it is boring, frustrating, irrelevant, etc.; low
job expectations, or ‘alternative’ job expectations such as crime; and negative peer
pressure or a culture of low achievement.
As the world changes even the promise of a good job can be unconvincing and in
many European countries there are university graduates with low grade jobs or no
job at all. The Education Accelerator must find new ways of motivating people to
learn.
Another point of relevance to the question of motivation is the creation of social
capital. There is a lot of research on the impact of social networking sites in creating
social and knowledge capital. These sites play an important role in helping people
raise their low self-esteem or become more integrated socially. So they may help create
appropriate social conditions for learning, raising individual expectations and giving
self confidence in pursuing higher educational and professional aims [77, 78].
ICT-enabled social networking is changing the way that people form groups al-
lowing individuals to transcend geographically local social structures and form social
networks outside the limited confines of the people in one’s neighbourhood. These
more virtual societies are often bound together by social structures that include the
dynamics of individual relationships [79].
There is a large and increasing body of knowledge of reputation systems, and this
will develop considerably through FuturICT over the next decade. This research will
be invaluable in helping to motivate students to learn using the Education Accelerator.
5.9 Games
Computer games are extremely popular with many people. A study of people playing
computer games in 2008 found the numbers of internet users playing computer games
once a month or more in the largest European countries were: Germany 29.3 million,
UK 23.8 million, France 20.2 million, Spain 14.6 million and Italy 12.9 million.
The reasons such large numbers of people spend so much time and effort play-
ing computer games include personal challenge through competition and a sense of
Participatory Science and Computing for Our Complex World 235
achievement at doing something well. It has been found that this motivation can carry
through when gaming techniques are applied in education. One such experiment is
teaching English in the context of computer games. The results so far suggest that
students find this approach to language learning much more attractive and motivating
than conventional methods [80].
Derryberry writes [81] that “Serious games [also called immersive learning simula-
tions, digital game-based learning, gaming simulation] are designed with the intention
of improving some specific aspect of learning, and players come to serious games with
that expectation. Serious games are used in emergency services training, in military
training, in corporate education, in health care, and in many other sectors of society.
They can also be found at every level of education, at all kinds of schools and uni-
versities around the world. Game genre, complexity, and platforms are as varied as
those found in casual games. Play, an important contributor to human development,
maturation and learning, is a mandatory ingredient of serious games.”
In his paper The Logic of Failure [82] Dorner give an example of how computer
simulation can show experts and students how not understanding many subtle in-
teractions can cause catastrophic failures in complex systems. A 2004 study by de
Freitas and Levene, validated in consultation with experts, comprised a consulta-
tion exercise with tutors and learners who use games and simulations regularly in the
learning practice. The conclusions from that study include “The majority of those ex-
perts interviewed thought that simulations and games significantly improved learner
motivation” [83].
In another context it has been found that students can become highly motivated to
study technical things by competitive robotics. In particular the RoboCup movement
has the objective of creating a team of humanoid robots able to beat the world
champions soccer players by 2050 [84]. Many of the scientists and engineers who will
work on this challenge are children today, so RoboCup has set up an organisation
called RoboCupJunior which has competitions in robot soccer, robot dance and robot
rescue. This combination of cooperative teamwork, creativity, design and engineering
is highly motivating for young people [85, 86]. In particular, some young men with poor
educational records find the challenge very motivating and are prepared to engage
and learn things which otherwise they would not.
Presenting education through gaming and competitions can have a big impact
on student motivation, transforming dreary or boring lessons into exciting interac-
tions with other students within social groups [87]. This can have a big impact on
motivating students with the Education Accelerator.
Spaced Leaning only or normal teaching. Experimental subjects (n = 67) studied the
course for just one hour in Spaced Learning, whereas controls (n = 258) were taught
over four months. All subjects were then tested in the high-stakes, multiple-choice
GCSE examination. The experimental subjects’ test scores demonstrated substantial
learning exceeding random answers (guessing) at a high level of significance. Sur-
prisingly the experimental students’ test scores were not significantly different from
controls’ test scores, even though they only had a fraction of the instructional time.
To achieve these outcomes, the GCSE Biology course normally taking four months
had been compressed into an intense 20 minute instructional presentation, introduc-
ing two or three major concepts per minute. As in all Spaced Learning, instruction
was repeated three times (“stimuli burst”), spaced by two ten minute distracter ac-
tivities (“sufficient intervals of inactivity”). Remarkably, students appeared to adjust
to Spaced Learning’s very intense learning and exceptional speed of delivery. These
results suggest it is possible to increase the speed of learning radically using Spaced
Learning. Of course, further research is vital to explore these results in different con-
texts, with different subjects, ages and methodologies. Trans-disciplinary neuroscience
and educational research into the dynamics of Spaced Learning may also reveal fur-
ther key triggers leading to other valuable innovations in learning.
The vision for the Education Accelerator as expressed in the grand challenges involves
(i) accelerating learning by an order of magnitude, (ii) reducing the cost of education
by an order of magnitude, and (iii) showing how this can be done. To achieve this
requires an action programme or plan to make these dreams become a reality within
the ten year FuturICT horizon.
The central claim to be investigated over 10 years is that together, the
Complexity, Social and Computing Sciences provide an urgently needed transdiscipli-
nary language for making sense of educational systems. In close dialogue with educa-
tional theory and practice, and grounded in the emerging data science and learning
analytics paradigms, this work will translate into practical tools (both analytical and
computational) for researchers, practitioners and leaders; generative principles for re-
silient educational ecosystems; and innovation for radically scalable, yet personalised,
learner engagement and assessment.
Our strategic plan divides the 120 months (ten years) of FuturICT into four phases.
Whereas Phase I can be specific and detailed, Phase IV must remain general and
aspirational since we cannot know for sure what will happen in the longer term.
However the objectives of the grand challenges are very clear, and like seeing the first
footprint on the moon, we can know for certain if they have been achieved.
Phase I prepares for the phases ahead. To be successful, a strong FuturICT educa-
tion community must be forged from the many scientists interested in education and
the FuturICT mission. This can be achieved by the well established and highly effec-
tive methods used by European ‘coordination actions’ which cover “definition, organi-
zation, and management of joint or common initiatives” including “activities such as
the organisation of conferences, meetings, the performance of studies, exchange of per-
sonnel, the exchange and dissemination of good practices, setting up common informa-
tion systems and expert groups.” (http://cordis.europa.eu/fp6/instr ca.htm).
Participatory Science and Computing for Our Complex World 237
5
By disruptive technologies we mean new technologies that disrupt established ways of
doing things including, for example, technologies that incrementally improve some particular
thing but undermine the whole edifice by enabling new kinds of approaches and systems.
238 The European Physical Journal Special Topics
Work is already underway in Europe to create a Digital Campus using the internet
to support research and teaching on a virtual campus with virtual tutorials, virtual
laboratories, classrooms and lecture theatres. The Digital Campus will implement
the radically new strategies discussed above, forming partnerships with professional
organisations such as the Complex Systems Society. It will develop new methods for
peer-to-peer learning and teaching, and for peer-to-peer assessment. It will implement
increasingly sophisticated automated and semi-automated assessment. It will enable
everyone to benefit from ubiquitous low-cost learning, i.e. providing direct access at
any time and in any location to the different responses to the questions he or she
has in mind. In other words, ubiquitous learning can be very fast and very pleasant,
simply because it addresses our immediate questions and needs. By its nature the
Digital Campus can implement ubiquitous learning and open to everyone the route
towards high-quality low-cost personalized education.
A Complex Systems Digital Campus6 is already close to becoming a reality.
Organised according to the UNESCO UniTwin scheme, over fifty university rectors
principles, and vice-chancellors around the world have signed up their organisations
as founder members. This shows already that there is a desire to collaborate across
nations and regions, and to share resources, experience and know how. Combined with
FuturICT educational community, this network of researchers and educators can be a
powerful force to meet the grand challenges of the FururICT Education Accelerator.
education systems are not fit for purpose given the huge need for education in Europe
and worldwide over the next ten years. Even in Europe many of our citizens are pro-
vided with ineffective education leaving many young people with poor knowledge and
skills and poor prospects for work and a satisfying independent life. Today education
is expensive, often ineffective, and frequently people of all ages find formal learning
difficult, demanding and unpleasant.
FuturICT requires and allows us to rethink education. Its Knowledge Accelerator
must be complemented by an Education Accelerator delivering more effective and less
expensive learning, enabling societies worldwide to educate all their citizens within
the resources available. This involves three Grand Challenges: (1) Enable people to
learn orders of magnitude more effectively than they do today; (2) Enable people
to learn at orders of magnitude less cost than they do today, and (3) Demonstrate
success by exemplary interdisciplinary education in complexity science.
These ‘man-on-the-moon’ Grand Challenges will be addressed by new scientific
understanding of evolving knowledge and new theories of learning; personalised learn-
ing and teaching; automated assessment of student progress and achievement; vir-
tual classrooms, lecture theatres and laboratories; social networking and peer to peer
teaching and learning; new ICT-enabled educational resources and new ways of shar-
ing of intellectual property; new ways of searching large education databases; and
new educational infrastructure such as the Complex Systems Digital Campus.
We have identified some of the most promising technological innovation strands
whose convergence with help tackle these Grand Challenges. A central claim to be
investigated over ten years is that together, FuturICT’s unique combination of
Complexity, Social and Computing Sciences could provide an urgently needed trans-
disciplinary language for making sense of educational systems. In close dialogue with
educational theory and practice, and grounded in the emerging data science and
learning analytics paradigms, this will translate into practical tools (both analytical
and computational) for researchers, practitioners and leaders; generative principles
for resilient educational ecosystems; and innovation for radically scalable, yet person-
alised, learner engagement and assessment. The proposed Education Accelerator will
serve as a ‘wind tunnel’ for testing these ideas, with an international virtual cam-
pus exploiting, testing and demonstrating the new understanding of complex, social,
computationally enhanced organisations developed by FuturICT.
References
89. R.D. Fields, Activity-dependent myelination, edited by R.D. Fields, Beyond the Synapse:
Cell-Cell Signaling in Synaptic Plasticity (Cambridge University Press Cambridge, UK,
2008)
90. R.D. Fields, The Other Brain: The Scientific and Medical Breakthroughs that will Heal
our Brains and Revolutionize our Health (Simon and Schuster New York, 2009)
91. P. Kelley, Making Minds What’s wrong with education- and what should we do about it?
(Routledge, London, New York, 2008)
92. P. Kelley, D. Reynolds, T.S. Whatson, ‘Spaced Learning: neuroscientific temporal codes
for memory acquisition in education’ (submitted) (2012)
93. K. Robinson, ‘Changing Education Paradigms’, RSA Animate, The Royal Society of
Arts, London, http://www.youtube.com/watch?v=zDZFcDGpL4U (2012)
94. R.S.J.D. Baker, K. Yacef, J. Educ. Data Mining 1 (2009)