BIOE 313: Neuromorphics: Brains in Silicon (EE 304)
Neuromorphic systems run perceptual, cognitive and motor tasks in real-time on a network of highly interconnected nonlinear units. To maximize density and minimize energy, these units--like the brain's neurons--are heterogeneous and stochastic. The first half of the course covers learning algorithms that automatically synthesize network configurations to perform a desired computation on a given heterogeneous neural substrate. The second half of the course surveys system-on-a-chip architectures that efficiently realize highly interconnected networks and mixed analog-digital circuit designs that implement area and energy-efficient nonlinear units. Prerequisites: EE102A and EE108 are required; EE114 is recommended.
Terms: Spr
|
Units: 3
|
Grading: Letter (ABCD/NP)
Instructors:
Boahen, K. (PI)
Filter Results: