Plastic mechanisms, multiple timescales and lifetime adaptation 

Guest Editor Ezequiel A. Di Paolo Evolutionary and Adaptive Systems, School of Cognitive and Computing Sciences, University of Sussex, Brighton, BN1 9QH, UK ezequiel (AT) cogs.susx.ac.uk

Editor-in-Chief Peter M. Todd Center for Adaptive Behavior and Cognition, Max Planck Institute for Human Development, Lentzealle 94, D-14195 Berlin, Germany editor@adaptive-behavior.org

The last few years have seen an increased interest in the design of plastic robot controllers, or controllers with inherent dynamical properties such as the interplay of multiple timescales, for the generation highly adaptive and robust behaviour. This research area in robotics draws important inspiration from neuroscience and may be applied to the testing and generation of hypotheses on the role of plasticity in brain function. Synthetic methods, such as evolutionary robotics, have provided a glimpse of how plastic neural mechanisms, like activity-dependent neuromodulation, that are often studied locally in reduced systems, can give rise to integrated and coordinated performance in a whole situated robot.

Recent studies have included the role of modulatory processes affecting neural activation, diffusing localized neuromodulation, the evolution of rules of synaptic change, the design of neural controllers acting on fast and slow timescales, and the evolution of stabilizing mechanisms of cellular activity. These studies have successfully revealed that such mechanisms are able to introduce highly desirable properties such as robustness, adaptation to bodily perturbations, and improved evolvability. But many questions remain open, such as what is the relation between plasticity and stability, how adequate is a given mechanism for the required task, how do alternative methods of obtaining plastic behaviour relate, and to what extent is environmental regularity responsible for successful tuning of neural controllers.

Adaptive Behavior solicits high quality contributions on these topics for its 2002 special issue (vol 10:3/4). Papers should describe work integrating mechanisms and adaptation at the behavioural level. They may present work using simulations or real platforms. Appropriate contributions addressing other levels of plasticity (such as sensory morphology or bodily structure) will also be considered. Papers drawing inspiration from, and contributing back to, neuroscience will be particularly appropriate.

Topics:

• Multi-timescale controllers
• Activity-dependent plastic neural controllers
• Change and stability in robot performance
• Adaptation to radical perturbations
• Neuromodulation
• Re-configurable neural controllers
• Plastic controllers and simulation/robot transfer

 

 

  Contents 

• E. A. Di Paolo (2002) Editorial. Adaptive Behavior 10 3/4 141-142.

• W. H. Alexander and O. Sporns (2002) An embodied model of learning, plasticity, and reward. Adaptive Behavior 10 3/4 143-159.

• T. Smith, P. Husbands, A. Philippides and M. O'Shea (2002) Neuronal plasticity and temporal adaptivity: GasNet robot control networks. Adaptive Behavior 10 3/4 161-183.

• T. Ziemke and M. Thieme (2002) Neuromodulation of reactive sensorimotor mappings as a short-term memory mechanism in delayed response tasks. Adaptive Behavior 10 3/4 185-199.

• E. Tuci, M. Quinn and I. Harvey (2002) An Evolutionary ecological approach to the study of learning behavior using a robot-based model. Adaptive Behavior 10 3/4 201-221.

• M. Lungarella and L. Berthouze (2002)On the interplay between morphological, neural, and environmental dynamics: A robotic case study. Adaptive Behavior 10 3/4 223-241.

• E. A. Di Paolo (2002) Spike-timing dependent plasticity for evolved robots. Adaptive Behavior 10 3/4 243-263.

• A. Ishiguro, A. Fujii, and P. Eggenberger Hotz (2003) Neuromodulated control of bipedal locomotion using a polymorphic CPG circuit. Adaptive Behavior 11 1 7-17.

• N. A. Borghese and A. Calvi (2003) Learning to maintain the upright posture: What can be learned using adaptive neural network models? Adaptive Behavior 11 1 19-35.

  Adaptive Behavior

 

Adaptive Behavior is the premier international journal for research on adaptive behaviour in animals and autonomous artificial systems. For over 10 years it has offered ethologists, psychologists, behavioural ecologists, computer scientists, and robotics researchers a forum for discussing new findings and comparing insights and approaches across disciplines. Adaptive Behavior explores mechanisms, organizational principles, and architectures for generating action in environments, as expressed in computational, physical, or mathematical models. Adaptive Behavior is published by Sage Publications, a leading independent publisher of behavioural journals, spanning human psychology to robotics to simulation modelling. A new editorial board, headed by Peter M. Todd of the Center for Adaptive Behavior and Cognition, is shaping the journal in novel directions. New technological infrastructure will allow faster publication turnaround, better feedback from reviewers to authors (and back again), and greater access to research results. The journal publishes articles, reviews, and short communications addressing topics including perception and motor control, learning and evolution, action selection and behavioural sequences, motivation and emotion, characterization of environments, collective and social behaviour, navigation, foraging, mate choice, and communication and signalling.