Professor Simon M. Lucas (SMIEEE)

School of Computer Science and Electronic Engineering

sml@essex.ac.uk ; tel: +44 1206 872048


IEEE T-CIAIG Image

Research Overview

My current research is focussed on AI and Games: prior to this I worked on pattern recognition, including various forms of N-Tuple classifer (these also turn out to be good for learning in games).

 

If you're interested in pursuing a PhD in AI and Games, check out our IGGI Centre for Doctoral Training

IGGI Logo

Current Teaching:

External Activities

 

Technical Committees

Chair (2006, -) of IEEE Computational Intelligence Society Technical Committee on Games

Chair of IAPR TC-5 Technical Committee on Benchmarking and Software (2002 - 2006)

Chair (2005) of IEEE Computational Intelligence Society Working Group on Evolutionary Computation and Games

Member of IEEE Neural Network Society task force on Co-evolution

Member of EPSRC Peer Review College (from 2006)
 

Journal

Conference Organisation

Competitions Chair, IEEE Congress on Evolutionary Computation 2009

Program Co-Chair, Parallel Problem Solving from Nature (PPSN) 2008

Competitions Chair, IEEE Congress on Evolutionary Computation 2007

Program Co-Chair, IEEE Symposium on Computational Intelligence and Games 2007

Competitions Chair, FUZZ-IEEE, July 2007, London, UK.

Publicity Co-Chair, Parallel Problem Solving from Nature (PPSN) 2006, Iceland

Program Chair, IEEE Congress on Evolutionary Computation (CEC 2006), part of IEEE World Congress on Computational Intelligence, May 2006, Vancouver

Finance chair for IEEE Congress on Evolutionary Computation
September 2005, Scotland

Competitions Chair, International Conference on Document
Analysis and Recognition
, Seoul, South Korea, August 2005

Competitions Chair, GECCO, June 2005

General Co-Chair, IEEE Symposium on Computational
Intelligence and Games
, Essex, UK, April 2005

Competitions Chair, GECCO, June 2004

Publicity chair for European Conference on Genetic
Programming
, Coimbra, Portugal, April 2004

Finance chair for Congress on Evolutionary Computation, June 2004, Portland Oregon

Special Session Co-Chair on Evolutionary Computation and Games,
Congress on Evolutionary Computation, June 2004, Portland Oregon

Competitions co-chair for Congress on Evolutionary
Computation
Dec 2003, Canberra

Competitions chair for International
Conference on Document Analysis and Recognition
2003, Edinburgh

Competitions co-chair for IEEE World Congress on
Computational Intelligence
,
May 2002, Hawaii

Competitions co-chair for Congress on Evolutionary
Computation
May 2001, Seoul, South Korea

Organiser and Chair of the first International colloquium
on grammatical inference (ICGI), University of Essex, 1993



Conference Committee Membership
 

Program Committee, Congress on Evolutionary Computation,
June 2004, Portland Oregon.

Program committee, European Conference on
Genetic Programming, 2004

Program committee, Congress on Evolutionary
Computation, Canberra 2003

Program committee, International Conference on Document Analysis and Recognition 2003, Edinburgh.

Program committee, 1st IEEE symposium on
Evolving Neural Networks, San Antonio, Texas, 2000.

Program committee, International Workshop on
Frontiers in Handwriting Recognition (IWFHR) 7, Amsterdam 2000.

Organising committee, British Machine Vision Conference (BMVC), University of Essex, 1997.

Program committee and organising committee, IWFHR 5, University of Essex, 1996.

Program committee, ICGI-96, Montpelier, France

Program committee and  organising committee, ICGI-94, Alicante, Spain

Program committee and organising committee, IEE workshop on Natural Algorithms in Signal Processing, Danbury Park, November 1993

Current Projects

Autonomous Car Racing

The challenge is to develop a software controller to race a standard remote control car around a track, against another car (either human or computer controlled).  The software must do this by taking live video feed from a web-cam and sending the optimal control signals to the car.  Comes in two versions: overhead view and on-car view (done with a wireless web-cam taped to the top of the car!).  More details here.

To help with the computer vision aspects of this project, I've developed a very simple Java Real-Time Video kit.

See also the IEEE CIG 2007 Competitions and this video by Julian Togelius and Renzo de Nardi

Evolving Finite State Machines

We've developed new evolutionary algorithms for inferring Deterministic Finite Automata (DFA) and Finite State Transducers from samples of data. Learning DFA is a well known problem, and we're recently tested our method on the excellent Gowachin server. We find the evolutionary method outperforms all other methods (most of which are based on heuristic state merging algorithms) when learning small machines (e.g. 20 states) from sparse noisy data samples.  This is work I've been doing with Jeff Reynolds.

Robust Reading

This is an old challenge that has proven remarkably tough. Modern OCR packages perform brilliantly on good quality documents but fail badly on poor quality images or general camera captured text. We've been developing a two-pronged attack on the problem, that involves the efficient application of contextual knowledge to optimally interpret the character hypothesis graph generated by applying a low-level classifier at all positions in the image.

For more information:

High Performance Sequence Recognition

I invented the scanning n-tuple classifier (SNT) in 1995 - a fast and accurate method for sequence recognition. This has since been investigated by other researchers world-wide. I've recently developed new discriminative training rules for the SNT, which typically outperform the conventional maximum likelihood estimation method.

Cellz

Games provide an ideal medium for exploring complex adaptive behaviour.  I'm currently working in a simple artificial life simulation game called Cellz.  This provides a minimal environment for studying principles of evolution and adaptive cooperative behaviour, with elements of competitive TSP (travelling salesman problem) etc.  More details here.