i++ Departmental Newsletter

Week commencing 10 December 2007

Previous Newsletters

News

AI and Games

The newly launched EPSRC Network on AI and Games got off to a fantastic start with its launch event at the Dana Centre, adjacent to Imperial College, London, on December 12.

The delegates included well known academic researchers and technical leaders from the games industry, from companies such as Electronic Arts, Eidos, Blitz, Lionhead, Sony, Introversion, and Rebellion.

Games have been transformed in recent years in terms of their graphical quality and more accurate physics models: the goal now is to achieve better AI. There were lively open-mic sessions, with well informed and passionate debate. These were used to drive round-table discussions, some of which may lead to research proposals and other forms of collaboration. There was general agreement that most game genres could be improved by better game AI, though the AI had to fit, and to work well; there’s no point having AI in a game just for the sake of it.

Pieter Spronck from the University of Maastricht gave an excellent keynote address richly illustrated with examples of recent successful games marred by the ridiculous behaviour of the non-player characters (NPCs). Simon Lucas is leading the NPC theme of the network, and will organise an event devoted to this at the University of Essex in 2008.

In the evening there was a public awareness event, also at the Dana Centre, including Dr. Lucas demonstrating his software agent Ms Pac-Man competition.

More details can be found here at the Dana Centre website.

Dr. Lucas commented: “it was wonderful to see so many members of the public with a keen interest in the latest developments in AI and Games.”
 

Staff News

Naci Balkan been elected to the Executive Committee of European Materials Research Society (E-MRS) for a period of three years, starting January 1 2008.

Founded in 1983, the European Materials Research Society (E-MRS) now has more than 3000 members from industry, government, academia and research laboratories, who meet regularly to debate recent technological developments of functional materials. The E-MRS differs from many single-discipline professional societies by encouraging scientists, engineers and research managers to exchange information on an interdisciplinary platform, and by recognizing professional and technical excellence by promoting awards for achievement from student to senior scientist level.

As an adhering body of the International Union of Materials Research Societies (IUMRS), the E-MRS enjoys and benefits from very close relationships with other Materials Research organizations elsewhere in Europe and around the world.

Each year, E-MRS organizes, co-organizes, sponsors or co-sponsors numerous scientific events and meetings. The major society conference, the E-MRS Spring Meeting, is organized every year in May or June and offers on average twenty topical symposia. It is widely recognized as being of the highest international significance and is the greatest of its kind in Europe with about 2000 attendees every year. Each symposium publishes its own proceedings that document the latest experimental and theoretical understanding of material growth and properties, the exploitation of new advanced processes, and the development of electronic devices that can benefit best from the outstanding physical properties of functional materials.

Papers Accepted

 "Optical and electrical properties of modulation-doped n and p type GaxIn1-xNyAs1-y/ GaAs quantum wells for 1.3 um laser applications"

Optical and Quantum Electronics

Naci Balkan

"Ubiquitous Robotics in Physical Human Action Recognition: A Comparison Between Dynamic ANNs and GP"

2008 IEEE International Conference on Robotics and Automation

 Theodoros Theodoridis, Alexandros Agapitos, Huosheng Hu, and Simon M. Lucas

Abstract
Two different classifier representations based on dynamic Artificial Neural Networks (ANNs) and Genetic Programming (GP) are being compared on a human action recognition task by an ubiquitous mobile robot. The classification methodologies used, process time series generated by an indoor ubiquitous 3D tracker which generates spatial points based on 23 reflectable markers attached on a human body. This investigation focuses mainly on class discrimination of normal and aggressive action recognition performed by an architecture which implements an interconnection between an ubiquitous 3D sensory tracker system and a mobile robot to perceive, process, and classify physical human actions. The 3D tracker and the robot are used as a perception-to-action architecture to process physical activities generated by human subjects. Both classifiers process the activity time series to eventually generate surveillance assessment reports by generating evaluation statistics indicating the classification accuracy of the actions recognized.

Seminar

Speaker: Dr Boris Mitavskiy, University of Sheffield
Host: Professor Riccardo Poli, Department of CES

18 January, 3.00pm in room 1N1.4.1
 

Title - Estimating the Stationary Distributions of Markov Chains Modelling Evolutionary Algorithms Using the “Quotient Construction” Method