Lately I have been giving much thought to the psychological applications of graph theory to neural networks. If indeed a neural network is an accurate model of the human brain, what may we discover by analyzing it as a graph? Is the brain a small-world graph? Does it have significant clumping? If its degree distribution, as I hypothesize, is not poisson, perhaps it isn't arbitrary after all, and what we thought was subjectivity is actually just the logic of pathways. But undoubtedly, since it is probably not a Erdos-Remyi graph, there's something there. The possibilities are limitless. And limiting: “although neural nets do solve a few toy problems, their powers of computation are so limited that I am surprised anyone takes them seriously as a general problem-solving tool.” Jeepers. I suddenly feel very small.
Subscribe to:
Post Comments (Atom)
Posts
No comments:
Post a Comment