Prerequisites: Analysis of algorithms, at the level of 6.046. Probability. Desirable: Distributed algorithms, 6.852 or a similar course. Some familiarity with system modeling and neuroscience.
Instructor: Prof. Nancy Lynch (email@example.com)
Schedule: TR11-12:30, room 32-124
Understanding computation in biological neural networks like the human brain is a central challenge of modern neuroscience and AI. This problem has been studied from many perspectives; this course will approach the problem using algorithmic methods from theoretical computer science.
Specifically the course will rely on synchronous, stochastic Spiking Neural Network (SNN) models. It will identify abstract problems to be solved by SNNs, including problems of focus and attention, similarity and clustering, representation, and learning. It will study algorithms (networks) that solve the problems, and analyze them in terms of costs such as network size and convergence time. Major emphasis will be on how concepts (both logical and physical) are represented in the brain, how these representations satisfy algorithmic goals, and how they are learned. Other emphasis will be on understanding how noise and uncertainty affect the costs of solving problems, and on how networks that solve simple problems can be combined into larger networks that solve more complex problems.