Professor Anthony Robins studies computational models of aspects of memory. What is the relationship between a solution to the "catastrophic forgetting" problem in neural networks, and the consolidation of learning in the sleeping brain? Do stable memories require stable synaptic changes? Can the properties of different kinds of state attractors help us to understand the nature of real memories, false memories, and category prototypes?
Associate Professor Alistair Knott and Dr Martin Takac are interested in models of working memory: specifically, models of the 'episodic buffer' proposed by Alan Baddeley, which is a short-term store of semantic material. Here's a paper describing a recent model.
M Takac and A Knott. A NEURAL NETWORK MODEL OF EPISODE REPRESENTATIONS IN WORKING MEMORY. Cognitive Computation, 2015.
