A few events recently have sparked memories of dreams I once had. Playing with Wolfram Alpha and seeing its error message “I am sorry Dave i’m afraid I can’t do that” indicating somewhat to its creators inspiration. My own inspirations and aspirations in entering the world of IT were very much the same. From the early attempts at simple conversational programs, programs that could play noughts and crosses and experimentations with Eliza it seemed like a small step to develop general purpose intelligence in computers if we just had a little more power. Prolog was a new hope at one stage. You could define the rules and the intelligence would work out how to solve the problem. Somehow it never seemed to get where we all hoped. We have computers now that will fit in a pocket with orders of magnitude more power than those that ran Eliza yet Wolfram Alpha seems to fall a little short of my inflated expectations.
I think the flaw is in our understanding of how we work, how our general intelligence solves problems. In Prolog all axioms needed to be specified, it is a closed world reasoning system. If it doesn’t know something then the answer is false. The human mind on the other hand is a generalising intelligence, as opposed to a generalised reasoning system. We can only reason about a limited set of things so we generalise our experiences and knowledge to fit those limited set of things and then reason about them. This is not a fool proof mechanism for interacting with the world but it is computationally efficient. Thinking back to the initial dreams of perhaps implementing HAL the key is in the name, HAL (Heuristically programmed ALgorithmic Computer) heuristics being simple rules of thumb that will get you an answer, not necessarily the right one but an answer and quickly.
Perhaps this is one of those projects which will make if off my “Someday Maybe” list eventually.
0 responses so far ↓
There are no comments yet...Kick things off by filling out the form below.