Bath AI Seminar Series: 2010


Bath Artificial Intelligence (BAI) Seminar is a research and discussion group. Announcements are made in the BAI mailing list, which also has its own archive.

Directions to Bath and a campus map: Computer Science is in 1W, beside the library.  Upcoming talks are listed below and announced by email.

Date Location Guest(s) Topic
10 Nov 2010 8 West 1.1 Michael Brooke
A jobbing scientist's random walk

Over the course of a career, a professional scientist can
become involved in many different and very diverse roles and tasks. These
can call upon many kinds of knowledge. The speaker will talk about his own
experiences in fields ranging from ion-exchange, through X-ray and
gamma-ray spectrometry and heraldry, to lip-reading using computers and
university teaching. This "random walk" and its implications will be
illustrated and recounted.

Lately Reader in Computing at the University of Bath and active in the
field of audio-visual speech processing since 1978, Dr Michael Brooke has
been a visiting scientist at the MRC Institute of Hearing Research in
Nottingham, at AT&T Bell Laboratories, Murray Hill, NJ and at the Speech
Unit of DRA Malvern. He was, with C. Benoît, a Co-Director of the NATO
Advanced Study Institute meeting on Speech-reading held at Bonas in 1995.
He has lectured widely in the USA and Europe and has also been secretary,
then chairman, of the Speech Group of the UK Institute of Acoustics. He
gained his Ph.D. as a physical chemist at Imperial College, London and has
also worked with the UKAEA at Aldermaston and at Lancaster University.
26 Oct 2010 1W 2.6 Joanna Bryson, University of Bath
The Role of Cognition in Cognitive Systems: From Robots to Primatology

The cognitive robotics community in Europe has been focused on an
approach known as developmental robotics, which (ironically)
de-emphasises the role played by the developer. While some argue for
this approach from a philosophical perspective, others admit that the
real motivation is that programming robots is hard.  My own approach,
Behaviour Oriented Design, focuses on making development of robots &
AI more generally as easy as possible. This includes getting the robot
to learn as much as it reliably can, but also means supporting the
developer and taking a systems approach to designing the entire
system. In this talk I describe the history of Behaviour Oriented
Design, including both robotic & scientific applications. I compare
this to contemporary scientific views of the interacting roles of
culture, cognition, genetic evolution and development.
13 Oct 2010 Wills Memorial Building,
University of Bristol
Dr. Andrew Hodges, Oxford University
Alan Turing: Computing for Life

Alan Turing (1912-1954) is conventionally counted as a British
mathematical logician, who in his own time was recognised for his
discovery of the fundamental property of computability in 1936. His
central part in the emergence of the practical electronic computer in
1945 was much less well appreciated by his contemporaries, and until
the lifting of official secrecy a generation later, very few knew of
his crucial codebreaking role in the Second World War. But from a
modern perspective, perhaps the most striking feature of his work is
that from the outset he was concerned with the nature of the human
mind and its embodiment in the brain. In this talk I will show how
this fundamental concern manifested itself in his early life and then
flowered in different ways through his computational innovations.

Dr. Andrew Hodges is a Tutorial Fellow in Mathematics, Wadham College,
University of Oxford. Andrew Hodges works primarily on new
developments in fundamental physics, but is best known to the general
public as the biographer of the British founder of modern computer
science, Alan Turing (1912-1954).

His book “Alan Turing: the enigma; has been translated into
several languages (winning a literary prize), has been dramatised for stage
and television and was chosen by Michael Holroyd as part of a list of
50 'essential' books (that were currently available in print) in The
Guardian, 1 June 2002. Dr. Hodges serves on the Turing Centenary
Advisory Committee.
18 May 2010 1WN3.10 Sam Brown (Zoology, Oxford)
Bacterial models of memetics

Microbes engage in a remarkable array of social behaviors, secreting
shared proteins that are essential for foraging, shelter, microbial
warfare, and virulence. These proteins are costly, rendering
populations of cooperators vulnerable to exploitation by nonproducing
cheaters arising by gene loss or migration. In such conditions, how
can cooperation persist? I review a series of models and data
highlighting the importance of population structure in determining
microbial social behaviours and consequent virulence, and highlight
the peculiar role and importance of microbial mobile elements
(plasmids, phages) in generating this structure. Finally, I discuss
the possible therapeutic implications of understanding the social
dynamics of microbial pathogens.
23 April 2010 8W 1.28 Alistair Gill, University of Surrey
Dr. Gill is a Research Fellow at the Centre for Research in Social Simulation
working with Nigel Gilbert. His research focuses on text-based communication
in online media with an emphasis on the social information that enables us to
understand others through text. In his talk he will describe the ability of
online users to accurately perceive others' personality, emotion and
trustworthiness despite the lean text-based format of communication. He will
also cover the caveats of these results. For example, people's ability to
interpret others' emotions and personality traits is more accurate depending
on the emotion or personality trait being expressed.

This seminar is pertinent to those working on computer-mediated communication,
collaborative systems and agent-human interactions.
09 March 2010 1E 3.6 Nick Hawes, University of Birmingham
Giving Robots Something To Think About

Many of the tasks we would like intelligent robots to perform involve
deliberation, i.e. explicit reasoning about future possibilities for sensing,
action, processing etc. In AI, planning provides a mechanism for such
deliberation, but typically requires general-purpose representations that are
at odds with the task-specific representations (e.g. those adopted by the
vision and language communities) commonly used to allow robots to sense and
act in the real world. In this talk I will present a framework which allows
planning to sit on top of a multiple, concurrently active, task-specific
subsystems. The framework has been tested on two different robots performing
different tasks (tabletop manipulation and human-augmented mapping), and
forms part of an information-processing architecture for intelligent robots
being studied as part of an EU integrated systems project.
 25 Feb 2010 8W 1.34 Hans Tompits, Technical University of Vienna
Methods and Methodologies (M&Ms) for Developing Answer Set Programs:
 project overview and initial results
17 Feb 2010 1E 3.6 Joanna Bryson, University of Bath
Modularity & Cognition

Time for AI  (about emotions, drives & priorities in action selection)
What limits the evolution of culture (biology & social science agent-based models)
Modularity in Cognitive Systems (about building cognitive systems)

AI at Bath