Computational Neuroscience course offered this semester

Published Date: 
January 7, 2019

Computational Neuroscience
Prof. Haim Sompolinsky, Hebrew University/Harvard

HARVARD GSAS: MCB 131 / NEURO 131
(cross-listed in Physics and SEAS)
canvas.harvard.edu/courses/49249
Mondays and Wednesdays, 3PM-4:15PM in Northwest B104

Questions? Email: Haozhe Shan: hshan [at] g [dot] harvard [dot] edu
Nimrod Shaham: nshaham [at] fas [dot] harvard [dot] edu
Haim Sompolinsky: haim [at] fiz [dot] huji [dot] ac [dot] il

Description: Follows trends in modern brain theory, focusing on local neuronal circuits
and deep architectures. Explores the relation between network structure, dynamics, and
function. Introduces tools from information theory, dynamical systems, statistics, and
learning theory in the study of experience-dependent neural computation. Specific topics
include: computational principles of early sensory systems; unsupervised, supervised and
reinforcement learning; attractor computation and memory in recurrent cortical circuits;
noise, chaos, and coding in neuronal systems; learning and computation in deep networks
in the brain and in AI systems.

Prerequisite: Basic knowledge of multivariate calculus, differential equations, linear
algebra, and elementary probability theory. This course is aimed at graduate students and
advanced undergraduates.