We may "bootstrap" a strong AI from a kind of knowledge database. After this initial bootstrap, does this AI remain unchanged?
"Real" strong artificial intelligence implies ability of discovering acceptable solutions from past experience when new encountered situations are partially or completely unknown.
How to evaluate acceptability of solutions?
A comparison between expected computed consequences and real events gives an evaluation. The result of this comparison is new knowledge itself.
Human beings constantly increase their knowledge this way. Artificial Intelligence shall use self-learning in order to think in the same way as humans do.
Real artificial intelligence = getting smarter from cumulated experience.
Once implemented such a property, is it compulsory to "bootstrap" a huge amount of knowledge?
Self-learning capability is a key feature for strong AI.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment