Towards implementation of a pattern recognition system based on a working theory of the learning rule for dendritic integration

My goal is to develop a working theory of the learning rule for dendritic integration by performing appropriate neurophysiological experimentation, and to then implement a pattern recognition system based on that learning algorithm so that the algorithm can be evaluated for its generalization ability. This thesis is a first step toward reaching that goal in that it presents an implementation of Jeff Hawkins and Dileep George's Hierarchical Temporal Memory (HTM) pattern recognition system based on an existing theory of the learning rule for dendritic integration -- spike-timing-dependent synaptic plasticity (STDP). I define probabilistic classification as the task of assigning a testing pattern a group of labels, where the pattern is classified correctly if its label is in that group, and classified incorrectly otherwise. I found that for visual pattern recognition, the STDP HTM system achieved much better overall probabilistic classification ability and far better generalization ability than the baseline HTM system.