Bayesian attractor neural network models of memory
Anders Sandberg · 2003
The work presented in this thesis deals with neural network models of human memory based on the Bayesian Confidence Propagation Neural Network (BCPNN). The focus is to explore how a model derived from a statistical framework can link more abstract top-down cognitive models with biologically plausible cortex models. Of special interest is whether it there exists necessary architectural differences between different memory systems or whether they can all be achieved within the same neural architecture for different parameter values.