David HORN eloadasa

Meghivo Az MTA Izotopkutato Intezet Adaptiv Rendszerek Osztalya meghivja Ont es munkatarsait 1996. aprilis 12-en, penteken 9h-kor tartando eloadasara. Az eloadas cime Eloado: David Horn Tel Aviv University Az eloadas kivonata: One of the paradigms of the field of Neural Computation is the Hopfield model of associative memory. This model is a generalized Ising model in which the binary spin variables represent neurons, and the coupling constants stand for synaptic weights which modulate the signals by which neurons communicate with one another. I will show how variants of this model can help us understand general features of the ailing brain, by employing such feedback neural networks to model memory systems of Alzheimer patients. A different paradigm is that of feed forward neural networks. I will discuss an example in which we use some information of how such networks function in the visual cortex in order to construct a system for particle identification in High Energy Physics detectors. The advantage of this system is that it performs parallel computations which may be implemented in hardware. Az eloadas helye: Konkoly-Thege ut 29-33, XXI/B epulet III. em. Tanacsterem. Lorincz Andras