There's a lot to like in this article. It looks at animal learning (including human learning from the perspective of a computer scientist, and then tries to apply the lessons learned back to artificial neural network (ANN) theory. Where the article gets interesting is where it looks deeply at just what constitutes learning, the role of experience in learning, and how (interestingly) experience shapes even innate knowledge through the long-term learning mechanism of evolution and natural selection. What are the lessons to be drawn? Pay more attention to network architecture, and look for underlying processes or skills that can be applied in multiple cognitive domains.
Today: 4 Total: 175 [Direct link] [Share]
