Introduction to Neural Networks
MIT OpenCourseWare offers this course on how networks of neurons compute and learn. It covers synaptic connectivity as the physical basis of neural computation, perceptrons, and the dynamical theory of recurrent networks, including amplifiers, attractors, and hybrid analog-digital computation. Backpropagation and Hebbian learning are treated as complementary models of how networks adjust their connections over time, and the course applies these frameworks to perception, motor control, memory, and neural development. Materials include lecture notes and problem sets from MIT's course, free to access under a Creative Commons license with no certificate offered. The content assumes some background in calculus and linear algebra and suits students moving from an introductory brain and cognitive science background into computational modeling of the nervous system.