The first day of the conference comprises a range of tutorials and a full-day workshop, to be held on Monday 13th December. Delegates will find these events to be especially valuable where there is a current need to consider the introduction of new AI technologies into their own organisations.
There are two streams of half-day tutorials, each with a morning and an afternoon session, plus an all-day workshop (the Ninth UK CBR Workshop). Delegates are free to choose any combination of morning and afternoon sessions to attend.
Professor Andrew Ware & Dr Hasan Al-Madfai, University of Glamorgan
The ability to predict some future event or condition is vital to many different spheres of human endeavour. These predictions are usually based on the results of study and analysis of available pertinent data. This tutorial will overview some of the statistical techniques that can help determine the saliency of data in making a prediction and investigate ways of making use of limited data. Moreover, the method by which these data can be most effectively used to build an intelligent forecasting device (typically a neural network) will be demonstrated. While the subject of intelligent forecasting will be covered in a generic way, use will be made of real-world examples throughout the tutorial.
Untangling the Semantic Web
Dr Alun Preece, University of Aberdeen
The Semantic Web is a relatively new concept, but already it means different things to different people. The World Wide Web community views it as a means to enrich the way information is interlinked on the global network. Knowledge engineers see the potential to automate problem-solving and information-seeking tasks. Logicians and ontologists are energised by the challenge of representing meaning and performing inference on a massive scale. This tutorial aims to untangle these various perspectives on the Semantic Web, while demonstrating that they are essentially compatible. Coverage will include both fundamental concepts and practical applications. No prior knowledge of the Semantic Web is required, although a familiarity with established Web technology will be assumed. Attendees should gain an understanding of the various viewpoints on the Semantic Web, and come away with knowledge of how to begin using – and contributing to – the next generation of the World Wide Web.
Recent Developments in Reinforcement Learning
Dr Eduardo Alonso, City University
Reinforcement learning describes the technique of learning through trial actions, whose reward values are updated based on the response of the environment. The key problem for reinforcement learning is how a software system can learn an optimal policy for balancing exploration and exploitation. The applications of the techniques designed to solve this problem are wide-ranging and include robotics, industrial manufacturing, and combinatorial search problems such as computer game playing. The aim of the tutorial is to provide an overview of such techniques, namely, dynamic programming, Monte Carlo methods, and temporal-difference techniques. The tutorial will explain how different algorithms work under various assumptions and how convergence and generalisation issues are addressed in both single-agent and multi-agent scenarios. It will finish with a survey of new approaches to the problem such as the combination of planning and learning, relational reinforcement learning, and animal-based reinforcement learning.
An Introduction to the Constraint Paradigm
Dr Marc van Dongen & Dr Barry O’Sullivan, University College Cork
Constraints are a declarative tool for specifying, representing, solving, and reasoning about many interesting problems occurring in computer science, mathematics, and – perhaps most importantly – the real world. Examples of successful applications of constraints include graphics (scene analysis, GUIs, etc); operations research (planning, scheduling, and other optimisation problems); molecular biology (DNA sequencing, constructing phylogenetic trees, protein folding, etc); electrical engineering (fault location in circuits, circuit layout, testing and verification, etc); mathematics (solving polynomial equations with precision, solving/simplifying equations over algebraic structures, etc); languages (parser construction for natural language, type checking for computer languages, etc); and much more. The tutorial will provide a basic understanding of how constraints can be used for problem solving. It will aim to cover constraint satisfaction problems (CSPs), consistency, search techniques, constraint representation, complexity, constraint paradigms, languages (OPL, ECLiPSe, OZ, CHR, CLP(R)), web resources, practical issues, and demonstrations.