9.012 BCS II, Part II: Math Background & Neural Networks

Spring 2001: M 2/16, W 2/21, F 2/23, M 2/26, 10-12 in 14-0637

Prof. Sebastian Seung, seung@mit.edu
Office hours: M 3:30-4, E25-210

T.A. Camillo Padoa Schioppa, camillo@mit.edu

Lecture schedule

Lecture 1 ( M 2/16): MATLAB and perceptrons. The abridged MNIST database.

Lecture 2 (W 2/21): Simple perceptrons

Lecture 3 ( F 2/23): Multilayer perceptrons

Lecture 4 (M 2/26): Attractor neural networks

Study questions

• What is the difference between the way in which feedforward and recurrent network models store memories?
• What is the evidence for multistability in the brain, and what does this phenomenon have to do with memory?

Readings

• J. Hertz, A. Krogh, and R. G. Palmer. Introduction to the Theory of Neural Computation, Addison-Wesley, 1991. sections 2.1-2.2, 5.1-5.3, 6.1, 6.3.
• J. Hopfield. Neural networks and physical systems with emergent collective computational abilities. Proc. Natl. Acad. Sci.USA. 79:2554-8 (1982)

Optional Recommended Readings

• David A. Robinson. Implications of neural networks for how we think about brain function. Behav. Brain Sci. 15, 644-55 (1992).