2005-04-27

Doubts creeping in

Today is a sad day.

I have doubts.

I have doubts that the simplicity of the predictor is a good thing.

Maybe I'm committing the same error AI researchers (the computer guys) committed all the time.

I'm ignoring the intricacies of the human brain (the six-layered details of the cortex).

Maybe it's important not to ignore it.

Why I have doubts?

There are several issues which are disturbing with the simple predictor:
- how it handles predictions coming from higher levels when there are several contradictory predictions from several higher level layers
- how it handles partial matches in input
- how it generates its own predictions based on partial histories

These things are getting out of hand.

It might be the case that extending the prediction/expectation part to handle more than one expectation and a list of candidates will solve the thing.

On the other hand it might not.

I have doubts.

Today is a sad day.

0 Comments:

Post a Comment

Links to this post:

Create a Link

<< Home