Adaption, learning, and plasticity in simulated neural networks.
Dr. Munro’s research addresses several aspects of neural networks, including mathematical properties, engineering applications, and models of neurobiological and cognitive phenomena. The majority of investigations have explored the capacity for adaptation and learning in these networks. Some specific projects have been the development of a scheme for concept formation, an account of the critical period for learning, a learning algorithm based on a scalar reinforcement signal, and a network for interpreting locative expressions. Current interests include an analysis of representations in multilayer networks, a model for learning spatial relationships, and the dependence of LTP and LTD on spike timing.