UMBC High Performance Computing Facility
Testing a computational model of biological neural
James T. Lo, Department of Mathematics and Statistics, UMBC
Bryce Carey, Department of Mathematics and Statistics, UMBC
David Alexander, Department of Mathematics and Statistics, UMBC
A computational model of neural networks was proposed that is a recurrent
network of processing units each comprising new models of dendritic trees,
synapses, spiking/nonspiking somas, unsupervised/supervised learning
mechanisms, and a maximal generalization scheme. The model shows how neural
networks encode, learn, memorize, recall and generalize.
In the project, we will test these capabilities and evaluate the model as a
learning machine for pattern clustering, detection, recognition and
localization on large benchmark data sets. Comparison with prior learning
machines will be performed.