Neural Networks 1 Class 4 19-4-2012

Herbrew University, ELSC-ICNC Ph.D. Program in Brain Research: Computation and Information Processing 76908 Theory of Neural Networks I Prof. Hanoch Gutfreund 4 credits 2nd Semester, Thursdays 9:00-12:00, Workshop: Tuesday 10:00-12:00 Brain Lecture Hall (Silberman Bldg., 3rd wing, 6th floor, Edmond J. Safra Campus, Givat Ram) The aim of this course is to provide students with the basic concepts of storage and processing of information in neural networks, and to equip them with analytical and numerical methods in the study and application of neural network models. In particular, the course focuses on concepts and methods of dynamics and their application to neural networks. The course syllabus includes: 1. Basic concepts: electrical properties of neurons, binary and analog neurons, deterministic and stochastic dynamics of the binary neuron, physical analogy — a spin in a magnetic field, networks of binary neurons. 2. Concepts from dynamics and statistical mechanics: master equation, dynamics of averages, detailed balance principle, thermodynamic equilibrium, energy, free energy and entropy. 3. Hopfield network: associative memory model, Hebb’s learning rules, Hopfield model, signal to noise ratio analysis, memory capacity, networks with asymmetric connections, the perceptron algorithm. 4. Networks of analog neurons: linear networks, firing rate models, nonlinear dynamics, bifurcation transitions, the Fitzhugh-Nagumo model, oscillations in excitatory-inhibitory networks
Video Rating: 5 / 5

This entry was posted in Neural Networks and tagged , , , . Bookmark the permalink.

Leave a Reply