Ep194 - Jeff Hawkins & Subutai Ahmad | A Thousand Brains: A New Theory of Intelligence
Talks at Google
00:00
Artificial General Intelligence
In the brain, we are not certain of the answer to something. We're trying to guess what's happening. A mathematician would represent uncertainty using perhaps a probability function. The brain doesn't work like that at all. It turns out, when you use sparsity, a sparse activations, the brain can represent multiplea hypothesis at the same time. So i'm using 200 neurons active to represent something out of ten thousand. Si've 200 active out of ten thousands, it turns out you can activate five or ten of patterns. And so there aren't proper cortical columns, but these cell types and ther circuit circuits exist in a different form than those found in humans. This is how nature
Transcript
Play full episode
