Internet of Learning-Things

Mass access to the Internet is a mere 20 years old and during this time Web Services have completely revolutionised how we interact – so how will the Internet transform us over the next 20 years?

This article explains how technologies can be architected to allow learning to flourish in the emerging world of the Internet of Things.

Beyond the “Internet of People”

In 2008, the number of things connected to the Internet exceeded the number of people on Earth – but that is still less than 1% of all the physical things in the world today. Cisco’s Internet Business Solutions Group (IBSG) predicts some 25 billion devices will be connected by 2015, and 50 billion by 2020, whilst IDC estimates machine-to-machine communication to grow to 41% of Internet communication by 2020.

IoT represents a major shift in how IT is being used. The personal computer and the ‘Internet of People’ defined the previous IT era. The Internet of Things will be defined by embedded and ubiquitous technologies such as 3d printing, advanced sensing and energy management.

A powerful illustration of this new world comes from wearable clothing, Tshirt OS from Cutecircuit –

Another is the rapid development and spread of 3d printing –

IoT is surging ahead in areas such as manufacturing, medicine and transportation, but what about education? ‘Smart Cities’ initiatives get plenty of attention, but what about Smart Schooling? What about an ‘Internet of Learning-Things’?

To help answer this question, eight schools in the UK will take part in a $1.2m scheme to find out how “Internet of Things” can enhance learning in science, technology, and geography. Students and teachers will be taught to measure and share data – using new Internet of Things technology – in ways that help make learning fun, link directly to the curriculum, and ultimately inform the design of the next generation of schools.

Whilst new-build schools in developed countries routinely use advanced energy and security management IoT technologies, a more fundamental shift is beginning to happen. There is a clear movement towards a Do It Yourself (DIY) approach to technology in the classroom. A great example of this is the such as the Bigshot digital camera kit –

A key part of this DIY trend is the increasing use of single-board miniature computers, particularly Arduino and Raspberry Pi. Arduino is a purely embedded system, while Raspberry Pi has both embedded and PC functionalities. Both are designed to teach computer science and electronics, and are optimized for managing control technology – i.e. the world of sensors, motors, displays etc (Things).

Floor Turtle and other technologies from the Constructivist movement have been around even longer. However, Arduino and Raspberry Pi have accelerated the Constructivist approach. To get results from these systems, users have to really understand how technology works, and once children understand the basics, their imaginations and creativity are unleashed. In an age when some ‘children think that cheese grows on plants’ one wonders where they think their consumer electronics come from, so its wonderful to see children becoming increasingly connected to the real world of how things work.

Arduino – the worlds’ most popular learning tool for electronics

Arduino and Rasberry Pi are surrounded by an extensive and complex ecosystem of devices and code, and one of the most noticeable devices is Makey Makey. Coming from the same stable that gave us Lego Mindstorms and Scratch, MaKey MaKey is a circuit-board with crocodile clips and wires which allow users to turn practically any object into a key from a computer keyboard. For example, a banana could be used for the letter ‘A’, some plastercine for the letter ‘B’, and a coin for the letter ‘C’. Using this simple principal, a staircase can be turned into a piano, or graphite pencil marks on paper could be used as a game controller.

Neither the Arduino or Raspberry Pi are anywhere near as prolific as PCs or Tablets, and they sell at a tiny fraction of the volume of the consumer and business devices that find their way into Education – tens of thousands a month as opposed to millions. However, unlike consumer and business PCs and Tablets, Arduino and Raspberry Pi have been designed specifically for education – so do they point the way forward?

The cost of a complete class set of Raspberry Pis (around $35 each) with Internet browsing, productivity tools, peripherals, sensors and devices would cost about ½ that of the equivalent class set of Tablets or PCs. However, the big drawback with Raspberry Pi is that they require patience and high levels of technical competency for their setup and operation – users need to become familiar with Linux and command-line prompts. At present the support ecosystem for Raspberry Pi is less than optimally organized for mass proliferation.

To get a better look at what the Internet of Things can mean for Education, we need to look beyond the ‘DIY’ world and think about a complete architecture for “Internet of Learning-Things”.

Towards an “Internet of Learning-Things”

Needs should drive the design of an Internet of Learning-Things – not the other way round. As with all questions about technology, the first question we need to ask is ‘why’? What new scenarios should an ‘Internet of Learning-Things’ deliver? Here’s some examples:

Technology literacy.

In the next 20 years machines will take increasing amounts of decisions. In a world where so much can be sensed or observed, security and privacy take on new meanings and relevance. In a world where systems will be managed increasingly remotely, technocrats will control much more of the world we live in. Its critical, therefore, that children get to understand how this completely new world works, and learn how to build and control it. To achieve this understanding, children need to have the opportunity to build systems that combine computer science with electronics and product design.

Science, Technology and Geography.

The use of sensors, data-logging and basic electronics has long been a part of the UK National curriculum, but with a proliferation of low-cost sensors, devices, drones and kits, its reasonable to expect to see an increase in the increasing use and sophistication in the application of these technologies across the world.

For example, the Parrot AR.Drone2.0 enables students to survey an area using a mobile phone. HD video is shot and stored on a USB memory stick, or relayed directly back to the phone. In one package, Science (e.g. physics of flight); Technology (e.g. OS, networking, control); and Geography (e.g. surveys, observations) can be delivered, in a way that is completely engaging for children of all ages.

The key development in this space is the opportunity for children to learn how to code with Scratch, Python and .NET Gadgeteer offering progressive learning pathway. Scratch even has a way to control the GPIO on Raspberry Pi, enabling students to control a range of devices easily.

Internet of Learning-Things - beginnings
Scratch offers an easy entry into the world of programming

Ubiquitous and context-aware learning.

With devices able to talk more easily with other devices, augmented reality should spill out from museums turning everyday features in the environment into learning objects. For example, point your phone at a building and see what was there of historical significance in the past; point it at a plant or animal and get key scientific facts; use a phone to control a drone and receive live images of your local neighborhood. Kiosks offer another platform for AR, and Lego have a powerful illustration that shows the kinds of scenarios that AR offers –

Learning through everyday play

A market research study by Tangull America indicated that the market for toys with embedded IT is growing over 15% annually, and will grow to sales of US $146 billion by 2015. Examples include interactive puppets, girls’ toys that share secrets, and “real playmates” – which measure changes in facial expressions and use AI to respond. There are huge opportunities to embed learning tools into children’s toys.

Personalised learning

With a greater spectrum of learning opportunities available, and wider use of project-based learning, the potential for more personalized learning increases.

Devices connecting securely to big (and nano) data, content and SRM systems, can enable more and better e-learning services that dynamically adapt to learner’s needs as they evolve.

“The growth of devices connected to the Internet will give learners access to untold sources of authentic data in an environmentally friendly way.  Through their Internet connections on multiple devices, learners will collect these data and work with fellow learners and experts around the world to analyse, interpret and manipulate the information and so contribute in a meaningful way to the development of social and scientific understanding, Learning will become more contextualised, relevant and meaningful as a result.”

Dr Michelle Selinger, Director of Education Practice at Cisco Consulting Services

Anytime anywhere high-stake assessment and exams

Nearly everyone on the planet has sat or will sit an examination or another form of high stakes assessment. Device-level security, built on biometric systems such as facial recognition, offer ways to ensure honesty in exams. As well as local devices, routers could be potentially enabled for exam-standard security in designated ‘Examination Zones’.

Towards an Internet of Learning-Things Architecture

The first technical problem that needs to be solved is that every device on the Internet needs an IP address to communicate with other devices. Currently most Internet traffic runs on IPv4, which allows ‘only’ 4.3bn addresses. The current version – Ipv6 – allows 7.9 x 1028 times more addresses, but IPv6 and IPv4 are not interoperable, so the transition is not going to be immediate and smooth.

The next problem to be solved is the development of protocols for data, network, transport, sessions and applications. A lot of work is underway such as MQTT, a machine-to-machine/Internet of Things connectivity protocol, but as yet there are no real IoT standards – unlike the Internet of People, which uses protocols such as http (for hypertext), and XMPP (for IM, presence and chat).

So, achieving any form of architectural standardization for an Internet of Learning-Things is going to take some time.

However, in the meantime, there are concepts and scenarios that can help. One way to look at IoLT architecture is to split it into functional layers, and map existing technologies and services to those layers:

IoLT Arch

Internet of Learning Things Scenario

A student has learned something significant and has verified the learning through a series of low stakes e-assessments. The student now wants to get full credit for this learning through an accredited examination board (eg, University of Oceania Certificate of Secondary Education). The student finds an accredited ‘Examination Zone’ – a room or an area set up to written examination standard, and monitored for honesty. The student logs onto the examination system, which verifies the user through device level biometric security, then locks down the device to ensure no access to local resources. The student is presented with the questions and types or handwrites the answers. The device pushes an encrypted version of the student’s answers to an E-Exam-Ready Wi-Fi router, (gateway) which relays the data to servers, which also have device level security to verify the validity and security conditions of the student’s responses. From there, the examination response is assessed and credit given in due course, with an encrypted certificate sent back to the student.

Whilst this may seem far-fetched and problematic, it’s worth taking a few moments to compare the kind of advances that have been made in Internet and mobile finance and medicine. For example, diagnostics in medicine is light years ahead of ‘diagnostics’ in education. In an era when we allow sensors to be implanted in the human body to monitor and improve health in the most precise and targeted way, why do we insist that practically everyone on the planet sits down in silence and recall facts from memory on bits of paper in order to get recognition for what they have learned?

Despite phenomenal progress with e-assessment and e-examination in some countries, a recent incident at Kasetsart University in Thailand illustrates just how far other places have to go. Students there were pictured wearing makeshift paper ‘anti-cheating’ devices.

The wrong kind of innovation
Draw your own conclusions – but no conferring please.


“We need to be ready for a new pace of change in learning”, says Jim Wynn, Chief Education Officer at Promethean.

“We will depend upon the content to be organized in ways which do not hinder learning and also and I think crucially, content will have to reflect next generation pedagogues and not those that are designed for the technology of pencil and paper”.

Another key point made by Jim is that the ‘Do It Yourself’ approach is not going to work on its own universally. “There has to be a balance between explore-and-find-out and directed learning from a wise head”.

Within formal learning, a major challenge is going to be lack of technical capacity amongst the teaching workforce. In developing countries, where some teachers don’t even know what Facebook is, ‘DIY’ will be a real challenge. Teachers in this new world will need to be a lot more technically skilled than they are now, and that will be a significant challenge.

Another challenge is the inertia in the examinations systems, and the cascade effect that it has on schooling as a whole.

One of the biggest challenges of all, however, is the uneven distribution of Internet Access across the world. Whilst it’s fascinating to talk theoretically about the Internet of Learning-Things in the developed world, what happens to those who are left behind from even the Internet of People?

According to the International Telecommunications Union, 39% of the world is not using the Internet. 31% of the developing world, and 77% of the developed world are using the Internet.

Internet users 2012, C/O International Telecommunications Union

There are several initiatives aimed at attacking this problem from different angles. For example, there is potential for using old analog TV bands – VHF/UHF – to deliver Internet access, whilst Project Loon is about delivering Internet access via high altitude balloons.

The Internet of Learning-Things will require significant amounts of virtual teaming. For example, the UK schools ILT pilot project will be led by DISTANCE, a consortium which includes at least 8 organisations, including 3 universities. Interestingly, DISTANCE plans to create a digital information hub using Xively Cloud Services – a cloud platform that is purpose-built for the Internet of Things.

An Internet of Learning-Things may be a long way off for some. However, in the same way that online content is beginning to be a disruptive force in formal schooling in some parts of the world, a new era of ultra-low cost and increasingly connected devices, sensors, displays, security and control technologies, is surely going to accelerate change in a very positive direction.

What is Ubiquitous Learning?

There is a lot of talk at the moment about Ubiquitous Learning. But what exactly is it, should we care, and how should it be implemented? This article by Edutech Associates member Nick Fekos explores these questions.

Ubiquitous computing is a model of human computer interaction in which computer processing has been integrated fully into daily activities, and also integrated into objects with which we routinely interact.  A Ubiquitous Learning Environment enables learning at any time, at any place.

Imagine you are a high school physics teacher and you are teaching concepts like gravity, friction, velocity and inertia. In a classic learning environment, you would be in your classroom with your students at a preset school period. But what if you could teach these concepts by taking your students to a soccer game or baseball game –

Origins of UL  

Mark Weiser from the Xerox PARC Lab ‘fathered’ UL more the twenty years ago. He envisioned three computer waves: mainframes which were prevalent at the time, personal desktop computers which were just appearing, and ‘Ubiquitous’ computing (also known as ‘ubicomp’), as the future. This third step is often referred to as reaching a point where the user is not aware of the computer, whatever form it has taken, but focuses only learning and the related materials.

Weiser identified three types of computer devices:

  • Wearable
  • Handheld
  • Interactive Boards

And their main characteristics would be:

  • Helpers/Servants
  • Quite and Invisible
  • User not necessarily aware of their presence, just the interaction
  • Should not demand attention

Key characteristics of Ubiquitous Learning

The main characteristics of ubiquitous learning are: (Chen et al., 2002; Curtis et al., 2002)

  • Permanency: Learning materials are always available unless purposely deleted.
  • Accessibility:  Access from everywhere as personally required
  • Immediacy:  Wherever a student is, he/she can immediately access learning materials.
  • Interactivity: Online collaboration with teachers and/or peers (chat/blogs/forums)
  • Situated instructional Activities: Learning in context (on-site).
  • Adaptability: Getting the right information at the right place for the right student.

Pedagogical Basis of UL

The main pedagogical premise of Ubiquitous Learning is related to ‘situated learning’ (see J. Lave and E. Wenger, 1991) which is a general theory of knowledge acquisition that is based on the notion that ‘true’ learning occurs in the context of real life activities. In contrast, formal classroom learning implies knowledge abstraction and decontextualization. This abstraction may not be such a problem, but learning in context (as illustrated at the beginning of the article) can certainly improve learning (as does engaging learners in authentic tasks).

Another pedagogical premise of UL would be collaborative learning (involving social interaction), again undoubtedly improving the learning process.

UL in the Context of Today’s and Tomorrow’s Technology 

Today’s technology seems to be trending towards the actualization of the original UL concepts as described by Mark Weiser.  Two out of the four essential components have already been established, and two are just now appearing as described below.

1. Mobile Devices: powerful, personal mobile communication, processing and storage devices

The proliferation of personal mobile devices, starting from smart mobile phones and currently progressing to tablets, has created an important shift in the direction of innovation as an intrinsic aspect of technology.  Perhaps not yet widely apparent in terms of the potential, but the shift has happened and is irreversible.

We now have a hardware device (a tablet) that is highly ‘personal’, similarly to how personal a mobile phone is, but much more personal than a desktop pc or a laptop.

This computing device, although in exchange for a certain degree of ‘personalization’ compared to mobile phones, is able to powerfully communicate, store, process and access information. It has the mobility and autonomy of a mobile phone, but the processing power and screen of a computer, and so it is much more suitable for broader and more fundamental use. Importantly, it provides the opportunity to move away from an ‘Angry Birds’ takeover of mobile technology

2. Cloud Computing

Cloud platforms can now provide the server side ‘omnipresent’ aspects of UL.  Any system with UL characteristics would have to be fully cloud based so as to ensure reliability and seamless scalability.  If design and development is originally geared towards maximizing efficiency by keeping required cloud power low, ‘lean’ cloud applications can be developed that can then be scaled much more powerfully, thus enabling efficient and robust UL.

Intelligent Personal Agents/Knowledge Objects

Given that we now have widespread truly mobile hardware devices, the next step is intelligent personalized software.

In order to truly implement UL and make ‘real’ use of available hardware and software platforms, the implementation of a personal knowledge object/agent that is ‘intelligent’ is essential.  Using Artificial Intelligence Techniques, this object/agent would take part in a ‘learning network’ (i.e. learn automatically) and would contain a rule base from which to make decisions.

This knowledge object/agent would model the ‘learner’ and would be dynamic.  It would have attached processes that would implement functionality like the ability to interface with other objects like itself, or to other non-intelligent objects (e.g. Word document) or to other systems (e.g. SharePoint) or devices (e.g. a telescope).

This interface functionality would be implemented using standardized file formats and access languages, like HTML5, SQL, RDF and OWL which are available today. The latter two introduce the idea of semantic processing, moving beyond the ‘text’ level into concepts and conceptual organization schemes (Ontologies).  Once we move into the conceptual processing realm (Artificial Intelligence), then very important and exciting functionality, like knowledge inference (reasoning) can be provided, which will mark a true technological turning point.

In summary, this platform independent knowledge object/agent would be the main vehicle for implementing Ubiquitous Learning (as described above) as it would know:

  • Who you are
  • Where you are
  • What device you are using
  • Dynamic skills and ability profile
  • Whether it is night or day
  • What time its
  • Who is near you
  • What devices are near you

Although seemingly too ‘futuristic’, the proliferation of wearable online devices will further the implementation of UL. A good example is Google Glasses (, with many more on the way.

A specific example

A student carrying a tablet approaches a telescope at school. The telescope ‘broadcasts’ its availability to the tablet which then informs the student of this. If the student agrees, the tablet connects to the telescope and sends information through its intelligent personal learning agent about the student, for example age, class, learning profile, interests, past projects and so on.

The telescope then transfers information that is appropriate for that particular student about itself, what it can do, and perhaps showing on the tablet screen what it is seeing right now. Also, the telescope connects to a cloud astronomy app, or to the Microsoft World Wide telescope for added experience and information.

Finally the telescope proposes a small interactive game from which it can assess the student to see what has been learned or not, and then perhaps contacting a fellow student to join the game online.

One thing is certain: the students would enjoy this, and so learning and assessment will have been achieved. This of course would be part of a broader educational strategy that would include other forms of learning, including classic learning paradigms.

Many of the pieces of the UL puzzle are now starting to fall into place, as summarised in the diagram below:

Please contact for more information.

Artificial Intelligence in Schooling Systems

Q. “What do you give a hurt lemon?”

A. “Lemon aid”

Like me, you may have thought that the writer of this joke is a student. Actually, the joke writer in this case is Artificial Intelligence software – a “joke generator” called JAPE.

Artificial Intelligence (AI) has growing implications for schooling, and this article aims to set out some of AI’s main concepts, and explore how they can be applied to improving learning.

What is Artificial Intelligence?

Artificial Intelligence is a mature field in Computer Science that has delivered many innovations, for example:

  • Deep Blue, the chess program that beat Kasparov
  • “iRobot Roomba” automated vacuum cleaner, and “PackBot” used in Afghanistan and Iraq wars
  • Spam filters that use Machine Learning
  • Question answering systems that automatically answer factoid questions

AI is best known for aiming to reproduce human intelligence. The field was founded on the claim that intelligence can be simulated by a machine. Essentially AI is the design of systems that perceive their environment and take action that maximize their chances of success. AI addresses natural language processing, reasoning, knowledge, planning, learning, communication, perception and the ability to move and manipulate objects. AI is about many things including interacting with the real world; reasoning and planning; learning and adaptation.

Different Approaches

There are several approaches to AI including:

  • Building models of human cognition using psychology and cognitive science
  • The logical thought approach with emphasis on “correct” inference
  • Building rational “agents” –  a computing object that perceives and acts

Key areas of application of AI in education include:

  • Robotics
  • Simulations
  • Games
  • Expert systems
  • Intelligent tutoring systems
  • Search, question and answers

Key AI Concepts

An initial view of AI reveals a field that is deeply divided into seemingly unrelated subfields. Some of these sub-fields even appear contradictory. For example, Neural Network techniques are considered by some a better model of human reasoning than rule-based Expert Systems, so lets take a closer look at these two approaches.

Neural Networks

This approach mimics the human brain through the use of “nodes”, which resemble neurons. Neural Network technology – which uses layers of “input”, “hidden (process)” and “output” nodes – has been applied successfully to speech recognition, image analysis, adaptive control, games and robots. Most of neural networks are based on statistical estimations, classification optimization and control theory. Neural networks can be programmed to model the behavior of natural systems – e.g. responding to stimuli.

Expert Systems

Expert Systems emulate the decision-making ability of a human expert by reasoning about knowledge – as opposed to following the procedures set out by a software developer as is the case of conventional programming. An expert system is divided into three parts – a knowledge base; an inference engine; and a dialog interface to communicate with users.

Machine Learning

Neural Networks can be applied to the problem of Machine Learning – the design and development of algorithms that allow computers to evolve behaviors based on data from sensors, input devices, or databases. An important task in Machine Learning is pattern recognition, in which machines “learn” to automatically recognize complex patterns, and to make intelligent predictions.

In games which have concrete rules and multiple permutations – eg Chess – Machine Learning calculates the most likely outcomes of the game given the position on the board by playing simulated games into the future. In addition, pattern recognition enables the game to analyze the relative merits of different moves in the game, based on which ‘shapes’ were created by experts in historical games.

Intelligent Agents

An intelligent agent is a set of independent software tools linked with other applications and databases running within one or several computer environments. Agent based technology systems include a degree of autonomous problem-solving ability. The primary function of an intelligent agent is to help a user better use, manage, and interact with a system or application. Additionally, software agents, like human agents (for example, an administrative assistant), can be authorized to make decisions and perform certain tasks.

Coach Mike, is an Intelligent Agent used at the Boston Museum of Science. Coach Mike’s job is to help visitors at Robot Park, an interactive exhibit for computer programming. By tracking visitor interactions and through the use of animation, gestures, and synthesized speech, Coach Mike provides several forms of support that seek to improve the experiences of museum visitors. These include orientation tactics, exploration support, and problem solving guidance. Additional tactics use encouragement and humor to entice visitors to stay more deeply engaged. Preliminary analysis of interaction logs suggest that visitors can follow Coach Mike’s guidance and may be less prone to immediate disengagement.

Enhancing Learning

Herbert A. Simon, an AI pioneer, said – “If we understand the human mind, we begin to understand what we can do with educational technology.”

Human learning and reasoning is founded on multiple knowledge representations with different kinds of structures, such as trees, chains, dominance hierarchies, neighborhood graphs, and directed networks. From MIT Open Courseware (Image by Prof. Joshua Tenenbaum.)

With systems that can both “learn” and provide “expertise”, the implications of AI for schooling are profound. Whilst AI has potential for solving problems like optimal resourcing and improving operational performance, the strongest area for the application of AI in schooling is to make learning more effective.

AI in schooling can be traced back to 1967 when Logo was created. Since the introduction of Logo and “floor-bots” such as Turtles, ever more sophisticated robots along with associated control technologies such as Lego Mindstorms – have been used in schools. Products such as Focus Educational’s “BeeBot” is a recent addition to systems applying some of the principles of AI in a schooling environment.

AI in schooling is evolving in several different ways:

Question and Answer Systems (QA)

By 2020, we’ll be creating enough data for a stack of DVDs containing it to reach the moon and back three times! Regrettably, the quality of answers does not necessarily improve in proportional to the amount of information available. The current generation of search engines are essentially information retrieval systems providing a list of “hits” from which the user has to deduce the closest match. One of the goals of AI, therefore, is to enable more natural questioning resulting in better answers and related information.

The first QA systems were developed in the 1960s as natural-language interfaces to expert systems. Current QA systems first typically classify questions and then apply Natural Language Processing. Natural language ‘annotations’ describe content associated with ‘information segments’. An information segment is retrieved when its annotation matches an input question. A generating module then produces sentences – ‘candidate answers’. Finally, ‘answer extraction’ processes determine if the candidate answer does indeed answer the question.

The implications for QA systems in schooling are enormous and raise significant questions about the role of teachers, learning content and assessment.

Learning With Expert Systems

Imagine students being given the task of recognizing patterns on science laboratory slides and making correct classifications. By combining expert and pedagogic models we are able to exploit AI to “mash” both domain specific and more general learning principles into a rich learning experience. When classifying the slides, students will be not just presented with a “right or wrong” response, but their behavior will be refined through “machine understanding” of why the student is making their decisions. AI differs from more conventional computing approaches by being able to generate and handle both “feed-forward” and “feed-back”.

Intelligent Tutors

Taking this a step further are Intelligent Tutors. These record their interactions with students to better understand how to teach them. Computer tutors are capable of recording both longitudinal data, as well as data at a fine-time scale, such as mouse clicks and response time data. Using these interactions as a source of data to be mined provides a new view into understanding student learning processes.

Games and Simulations

Currently, the area in which AI is applied the most is Computer Games – and by a large margin. The use of scenario-based simulations and serious games for training has been well-accepted in many domains. Simulations require active processing and provide intrinsic feedback in an environment in which it is safe to make mistakes. Artificial ecosystems – like the one shown below – have proved popular and have their uses in schooling.

An interesting learning mechanism used in game based learning that is potentially usable in other contexts is “Transfer Learning” – which can help improve the speed and quality of learning. The idea is to use knowledge from previous experiences to improve the process of solving a new problem.

Two key AI methods underpin this approach –

  • Case-Based Reasoning (CBR) – a set of techniques for solving new problems from related solutions that were previously successful.
  • Reinforcement Learning (RL) – set of algorithms for solving problems using positive or negative feedback from the environment.

Reinforcement Learning can be delivered through the following mechanism –

  1. A central database with a collection of rules, mapping all possible actions and relative values.
  2. A learning component that takes feedback from the environment, and updates the utility value of each action. This is done using a reinforcement learning policy which estimates if there were any improvements since the last step.
  3. A planner then takes these rules, and computes a plan of action randomly based on the utility of the actions.

To anyone who has explored managed learning, this should sound quite familiar.

Two interesting models for understanding human learning in AI and Games context have come out of Microsoft Research:

This model classifies different types of learning in the context of games environments, but has transferability to broader understandings of the interface between computing and learning:

This model helps visualize the relative ease with which a game player can learn, depending on the granularity of detail presented to them:

  • Too coarse: cannot learn a good policy
  • Too fine: impossible to learn from little experience
  • Just right: learn a good policy from little experience

Personalized learning

Ramona Pierson, Chief Scientific Officer for Promethean, talks about ‘mashable’ digital content with embedded assessments tightly coupled to the curriculum, and learning progressions made ‘dynamic’ by AI. This can adjust learning progressions continually for each student, presenting cross-curriculum content and learning strategies based on a dynamic learning process.

“Imagine how powerful it would be for a student to have a customised textbook, sequencing of lessons, and embedded assessments that dynamically changed to ensure that he/she masters the material in the way that makes sense, and would result in obtaining nationally set benchmarks and learning outcomes”. (Mass Customisation And Personalisation Of Learning, Education Technology Solutions).

Nick Fekos, a former AI programmer in the financial sector and now at Athens College, agrees and is formulating plans for an intelligent object oriented knowledgebase that ‘learns’ from ‘experience’ and adjusts accordingly. The system Nick has in mind will implement dynamic, self-organizing and differentiated learning paths. The more the learning algorithm is used, the better it will get – perhaps something that can be said for the more general application of AI to schooling itself.

So How Do I Build an AI System?

Firstly, there is plenty of opportunities for getting students developing AI systems.

Besides Logo, its worth looking into Kodu – a  visual programming language made specifically to enable children to create games.

Also check out Microsoft Robotics Developer Studio which helps make it easy to develop robot applications. The current version (4), which is in Beta, provides extensive support for the Kinect sensor hardware allowing developers to create Kinect-enabled robots in both a ‘Visual Simulation Environment’ and real-life.Integrating AI into other learning workloads is an altogether more complex task.

For anyone wanting to understanding the mechanics of programming an AI system, this excellent article shows how to programme a neural network in C#.

For a more comprehensive desicription, including important architectural principles, check out this paper from University of Southern California which explains how to build a simulation to teach soft skills such as negotiation and cultural awareness.

For a comprehensive coverage of the field of AI in Education, look at the proceedings from Artificial Intelligence in Education, 15th International Conference, AIED 2011, Auckland, New Zealand, June 28 – July 2011.

For a comprehensive coverage of the field of Intelligent Tutoring, look at the proceedings from the 10th International Conference on Intelligent Tutoring Systems, ITS 2010, Pittsburgh, PA, USA, June 14-18, 2010