Swarm Intelligence from Particle to Human

I am due to give a talk at Le Havre Universitie tomorrow 24th Jan 2008 on Swarm Intelligence with video link to Rouen University. Here is the abstract, which I hope will initiate some discussions. Slides are to follow.
==============================
Vous êtes cordialement invité au
Prochain séminaire du LITIS (axe ISC)
Jeudi 24 Janvier 2008
14h Salle de visioconférence de l'UFRST du Havre
et Salle de
visioconférence de l'Université de Rouen - Madrillet


==============================

======================
Swarm Intelligence and Applications
Optimization, Nano Technology, Social Simulation, and Art
Par
Aladdin Ayesh,
Co-ordinator of Intelligent Mobile Robots and Creative Computing
Research Group, De Montfort University, UK




Swarm intelligence is the result of observing nature where we see
swarms in the form of bee hives, ant colonies, chemical structures,
physical particles. In all these cases, these small tiny entities
interact and work together producing an intelligent behavior from
simple rules.

In recent years, there has been an expansion in using swarm
intelligence in a number of ways. First, swarms used as a simulation
tool for sociological and ecological studies. This perhaps their first
real application. Second, a number of optimization algorithms based on
swarms emerged, e.g. Ant Colony Optimization and Particle Swarm
Optimization. Third, nano-technology advances increased the interest
in hardware swarms, i.e. nano-robots, and software swarms, e.g. to
simulate bio-chemical structures and their physical properties before
attempting to create them in the labs. Finally, swarms are social
structures that fascinated artists, e.g. cellular automata and pixel
hives.

This seminar will explore swarm intelligence as a current and active
trend in artificial intelligence community. A brief history and themes
of swarm intelligence will be covered, then the main application areas
will be discussed. Some on-going projects at De Montfort University
Intelligent Mobile Robots group will be discussed. The seminar will
conclude by identifying some areas of growth in swarm intelligence
especially those related to complex systems.
========================================================

Emotion Modeling for Reasoning and Interaction

I am due to give a talk at Le Havre University on Tuesday 15th, 2008 on the subject of emotions. As I was writing the abstract it evolved into a short article. Whilst a shorter version is used to publicize the talk, here is the extended abstract of the talk.
================================================

Emotions have been recognized as an important factor of the human composition since the early human thoughts and scientifically studied as early as 1872 with Darwin's book 'The Expression of the Emotions in Man and Animals'. This followed by a number of theories and schools, e.g. Lang (1880), Watson (1919), and Ekman (1978) to mention but few. However, early researchers working on formalizing knowledge and reasoning ignored the impact of emotions and depended on rational logical systems (e.g. propositional and predicate logics, modal logics, etc.).

Nowadays, cognitive processes have been understood to go beyond the logical reasoning. The connections between memory, perception, and desires are being recognized as important factor in human reasoning and decision making. These ideas filtered through to Artificial intelligence research, e.g. Belief-Desire-Intention (BDI, 1987). However this filtering remained within the boundaries of computational solution dressed into psychological and sociological analogies tagged as Folk Psychology.

The recent advances in cognitive psychology, e.g. the work of Anderson, Miller and Piaget, helped by advances in neuro psychology and advances in neurological computational models started to give stronger scientific basis to studying human cognition. As a result, a growing research community on cognitive systems evolved looking at scientific approaches to research cognition and to develop cognitive systems. This growing interest is also reflected in the EU and national research councils funding. Cognitive systems is now a main stream theme in EU FP7.

These advances in cognitive systems gave a greater importance to emotion modeling research. The more we try to make a cognitive agent, and often fail, the more that emotion effects are recognized. The relation between emotions, perception and action selection, especially in dynamic uncertain environments, have been the primary study of a number of projects, e.g. Flame project (2000). There are several approaches to modeling emotions. Cognitive maps, fuzzy logic, multi-valued logics and neural networks are few examples of techniques used to model emotions. The interesting aspect about these new modeling approaches is that they attempt to capture the essence of emotions beyond the early computational attempts that used emotions as mere elaborated thresholds.

In this seminar, we will explore the psychology of emotions, discuss emotion modeling for cognitive agents with particular focus on two areas. The first is the effects of emotions on reasoning that includes action selection, decision making and problem solving. The second focus will be the effects of emotions on social interaction between cognitive agents within rational or irrational situations, e.g. crowd simulation in normal and under pressure situations.

The talk will conclude with areas of interest that is still open to investigation. In particular, the effects of emotions on memory and association between experience, perception, emotions and evolving personality. The applications of emotional agents in games and other applications will be discussed.