SGAI

AI-2012 Thirty-second SGAI International Conference on Artificial Intelligence. CAMBRIDGE, ENGLAND 11-13 DECEMBER 2012

home | schedule | technical stream | application stream | poster sessions
workshops | proceedings | exhibition | registration | sponsors | organisers
enquiries | social | visa info | venue | accommodation | panel session
cbr tools tutorial | speakers | previous conferences |

call for papers | paper submission and info for authors | accepted papers
internet access for delegates

BCS

Workshops

The first day of the conference comprises a range of workshops, to be held on Tuesday 11th 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 will be four half-day workshops plus an all-day workshop, the Seventeenth UK CBR Workshop. Delegates are free to choose any combination of morning and afternoon sessions to attend. The programme of workshops is shown below. Note that the morning session starts at 11 a.m. to reduce the need for delegates to stay in Cambridge on the previous night. There is a lunch break from 12.30-13.15 and there are refreshment breaks from 14.45-15.15 and from 16.45-17.00.

Workshops organiser: Professor Adrian Hopgood, Sheffield Hallam University


Stream 1 - Morning (11.00-12.30 and 13.15-14.45 Upper Hall)

Practical Data Mining

Chair: Dr Peter Lane, University of Hertfordshire
Co-chair: Dr Na Helian, University of Hertfordshire

The standard data-mining / machine-learning module taught at university level introduces a range of techniques for classification or clustering, but does not typically have the time to cover all aspects of real-world data-mining such as how to create a feature set, which instances to use for training, and how to evaluate a model against its expected use. In addition, students are typically exposed to standard data-mining toolkits such as WEKA which are rather unwieldy to manage in complete applications through the graphical interface.

This workshop will help the participant bridge the gap from basic knowledge to effective practice by using some standard libraries and small programs written in the Ruby language. The workshop will work through all stages of a realistic data-mining application, introducing those practical techniques which often stand between disappointing results and success.

The aim of this workshop is to bridge the gap from basic knowledge to effective practice in the use and application of machine-learning and data-mining techniques. To achieve this, the workshop will:

  1. introduce the capabilities of Java-based libraries such as libsvm and WEKA;
  2. propose Ruby as a scripting language for these libraries;
  3. demonstrate good machine-learning practice and typical experimental setups;
  4. illustrate these with a realistic application in the area of text mining or image analysis, covering topics such as feature selection, instance selection and model evaluation.

The workshop should be of interest both to students beginning research projects with a learning component, and to commercial data analysts wishing to introduce data-mining techniques into their work.

You are invited to participate in some practical exercises, for which you will need a laptop with Java 6, a text editor, and a command-line environment. However, you will still benefit from the workshop if you don't have a suitable machine.

All participants will receive the following on a CD-ROM:

  • jRuby (see http://jruby.org)
  • Standard libraries
  • Rubygems (as needed, some written by the presenter)
  • Copies of slides covering the materials
  • Notes explaining the main techniques, including a bibliography with pointers to academic papers and other relevant literature.
  • Fully annotated example code

Stream 1 - Afternoon (15.15-16.45 and 17.00-18.30 Upper Hall)

Temporal Representation and Reasoning in AI

Chair: Dr Jixin Ma, University of Greenwich

We understand 'time' when we speak of it or hear it spoken of by another, but are likely to struggle if asked to explain it. Temporal representation and reasoning play a ubiquitous and vital role in computer science. In particular, many artificial intelligence systems need to deal with the temporal dimension of information, i.e. the changes in information over time and knowledge about how those changes occur. In fact, time seems to play the role of a common universal reference. Everything appears to be related by its temporal reference, although temporal references may have different forms, such as:

  • absolute temporal entities (e.g., "3.15 pm on 11th December 2012"), which refer to explicit time elements;
  • relative temporal entities (e.g., "during the time when the officer was in his office"), which refer to time elements that are known only by their relative temporal relations to other time elements that, again, may be absolute or relative;
  • absolute temporal durations (e.g., "45 minutes"), which refer to some certain amount of temporal granularity;
  • relative temporal durations (e.g., "less than 3 hours"), which refer to some uncertain amount of temporal granularity.

The problem of reasoning with temporal information in the mixture of these forms is twofold: how to represent various kinds of temporal knowledge and how to construct a reliable method of inference, based on the representation.

This workshop will motivate discussion on an important and interesting topic in artificial intelligence. The speaker will present an overview of some fundamental issues with respect to temporal theories and models. He will also introduce the three main approaches to the representation of temporal information. Some interesting examples in temporal representation and temporal reasoning will be introduced.


Stream 2 - Morning (11.00-12.30 and 13.15-14.45 Lubbock room)

Natural Language Processing

Chair: Dr Ariadne Tampion

This workshop will start by asking the questions 'What is language?' and 'Why do humans use it?' in the hope of uncovering some insights that will be of use when programming machines to do so. This will be followed by an introduction to language processing, concentrating on techniques which have actually made it into deployed systems. The next two presentations will showcase cutting-edge projects from academia and business. Finally there will be a discussion in which all attendees can participate. The workshop should offer something for everybody with an interest in the field: from the newcomer to the established practitioner.

Programme
11:00 Chair's Welcome and Introduction
11:10 Amorey Gethin, The English-Learning and Languages Review,
'A rational take on language and the contribution of Ronald Englefield'
11:50 Dr Paula Buttery, University of Cambridge,
'An introduction to language processing concentrating on techniques that have made it into deployed systems' 12:30 Lunch
13:15 Dr Andreas Vlachos, University of Cambridge,
'The SpaceBook project: Assisting tourists in navigating and exploring a city'
13:45 Chris Ezekiel, CEO, Creative Virtual
'Virtual assistants delivering real business benefits'
14:15 Discussion: The challenges of programming machines to 'talk like us'
14:45 Tea

Any queries regarding this workshop should be directed to the Chair via her website.

Stream 2 - Afternoon (15.15-16.45 and 17.00-18.30 Lubbock room)

Robotics and Intelligent Control

Chair: Dr Hassab Elgawi Osman, University of Tokyo

Intelligent control has been a long-standing goal, yet remains an intricate challenge in the science and engineering of control. Despite tremendous growth in computational technologies, today robotics and autonomous machines/systems are confined and less autonomous because our best control algorithms exhibit only little adaptability. On the other hand, reinforcing learning (RL), a machine learning (ML) paradigm, offers a myriad set of promising methods especially suited to learning control policies. Some have been already applied with relative success, able to reproduce some simple control tasks in artificial devices, both simulated and real.

This half-day workshop emphasises the use of wide range of strategies to build control algorithms in autonomous systems. This theme embraces aspects related to, but not limited to, design principles, sensing, automated reasoning and planning, SLAM (simultaneous location and mapping), multi-robot systems, etc.

Invited speakers will demonstrate how existing ML techniques can be developed further for solving more challenging control tasks. The long-standing goal of the Workshop Chair is to provide new design principles toward the development of intelligent machines that carry out fully autonomous tasks.

Programme
15:15 Chair's Welcome and Introduction
15:20 Prof. Mamoru Minami, Okayama University, Japan
           "Eye-Vergence Visual Servoing of Robotic Manipulator by Evolutionary Pose Estimation"
16:00 Prof. Trevor Martin, University of Bristol, UK
            "Verification of Autonomous Systems - The Role of Uncertainty"
16:45 Tea
17:00 Dr. Woon Jong Yoon, Qatar University, Qatar
           "Evolution: From No Intelligence to Autonomous and Smarter Biomedical Diagnostic and Assistive Devices"
17:45 Dr. Hassab Elgawi Osman, Visiting Researcher, University of Birmingham, UK
           "Promising ML Techniques in Robotics and Adaptive Control A short Survey"
18:00 Discussion
18:25 Closing & Announcements


Stream 3 - All Day (11.00-12.30, 13.15-14.45, 15.15-16.45 and 17.00-18.30 Peterhouse Lecture Theatre)

Seventeenth UK Case-Based Reasoning workshop

Chair: Professor Miltos Petridis, University of Brighton

Workshop Programme

SGAI

AI-2012 Thirty-second SGAI International Conference on Artificial Intelligence. CAMBRIDGE, ENGLAND 11-13 DECEMBER 2012

home | schedule | technical stream | application stream | poster sessions
workshops | proceedings | exhibition | registration | sponsors | organisers
enquiries | social | visa info | venue | accommodation | panel session
cbr tools tutorial | speakers | previous conferences |

call for papers | paper submission and info for authors | accepted papers
internet access for delegates

BCS