Associative Learning and Reinforcement Learning Workshop
Bristol, UK, 3rd April 2006
2nd CALL FOR PAPERS

The Associative Learning and Reinforcement Learning  workshop(http://www.soi.city.ac.uk/~eduardo/CAL-AISB.html) aims to encourage interdisciplinary research by strengthening collaboration between learning scientists and the artificial intelligence community. The workshop will be held as part of the annual convention of the Society for the Study of Artificial Intelligence and Simulation of Behaviour (AISB -- Bristol, UK, 3-6 April 2006).

Connectionist models have been consistently used in the study of associative learning, and standard models of classical conditioning can be naturally interpreted in terms of computational neuroscience.  Moreover, the links between classical conditioning and temporal-difference learning (one of the central elements of reinforcement learning) are close and widely acknowledged.   Despite this, important issues remain that indicate the distance between these two areas of research; for example, reinforcement learning does not take full advantage of current associative learning theories, whereas associative learning uses computational models as ad hoc tools rather than complementary formal models of behaviour.

The broad purpose of this symposium is to give a representative overview of current research in the area, in order to facilitate dialogue between the two scientific communities. We anticipate the potential benefits will be bidirectional: on the one hand, advances in associative learning may help improve the performance of existing reinforcement learning algorithms (e.g., in improving generalisation to different learning environments); on the other, intelligent systems designed with more psychologically-based constraints might provide interesting psychological insights (e.g.,  
into the mechanisms underlying goal-directed behaviour).  In addition, the session will provide an ideal scenario for the presentation of learning simulators and psychologically inspired adaptive robots.

The workshop will comprise invited talks by Anthony Dickinson (Experimental Psychology, Cambridge) and Nathaniel Daw (Gatsby Computational Neuroscience Unit, UCL), followed a number of shorter talks and an open discussion.  Authors are invited to submit either complete, or tentative and ongoing work by the 13th of January 2006 (to eduardo@soi.city.ac.uk).  Accepted submissions (either abstracts or papers) will be published by the AISB as on-the-day (copyright  
free) symposium proceedings. Formatting instructions will be available from the symposium website shortly. We are also negotiating with a prestigious journal the publication of an after-event Special  
Issue consisting of extended best papers, as well as some additional,  invited contributions.

Organisers:

Eduardo Alonso, Computer Science, City University
Charlotte Bonardi, Psychology, University of Nottingham
Esther Mondragón, Psychology, University College London

Programme Committee:

Christian Balkenius, Lund University Cognitive Science Bernard  
Balleine, Department of Psychology, UCLA
Joanna Bryson, Department of Computer Science, University of Bath
Peter Dayan, Gatsby Computational Neuroscience Unit, UCL
Magnus Enquist, Deparment of Zoology, Stockholm University
Geoffrey Hall, Department of Psychology, The University of York
Rob Honey, School of Psychology, Cardiff University
Robin Murphy, Department of Psychology, UCL
Ulrich Nehmzow, Department of Computer Science, University of Essex
Yael Niv, Gatsby Computational Neuroscience Unit, UCL
John Pearce, School of Psychology, Cardiff University
José Prados, School of Psychology, University of Leicester
Richard Sutton, Department of Computing Science, University of Alberta