CS 1501: AGI Lecture 8 | Neuromorphic Computing and Spiking Neural Networks

“What I see in Nature is a magnificent structure that we can comprehend only very imperfectly, and that must fill a thinking person with a feeling of humility. ”

― Albert Einstein 

Course Recap:

  1. ANNs, while a breakthrough, do not accurately model the way the brain learns and represents information.

  2. We still do not know for sure how the brain learns - through backpropagation, hebbian learning, or STDP.

  3. Spiking Neural Networks are an attempt to make ANNs more accurate, and work by passing discrete “spike trains” through the network, instead of continuous data input.

  4. SNNs, because of the differential equation that they use to model neurons, are expensive to simulate, so new hardware was introduced for them.

  5. This hardware, called neuromorphic computing, attempts to create hardware level neurons for SNNs to run on effectively.

  6. These might be an effective way to an AGI.

  7. Gazzaniga's declaration of determinism might be overcome with advanced neuromorphic chips or other architectures, like quantum computing. 

View Full Lecture PDF Below:

Week 8_ Neuromorphic Computing and Spiki
Download • 1.28MB

6 views0 comments