Ambiguity can reveal itself even in very simple sentences. Context can help clarify such ambiguities, but how? 1) "spreading activation" among specific agents, 2) sharpening of activation by recurrent connections. Problem: with this notion, won't everything at some point activate? Solution: no -- prevent this with cross-exclusion mechanisms.
(NB: This looks very much like a recurrent version of a semantic network. What kind of model? Well, the IAC network (for instance) that we demonstrated with the Sharks/Jets data set!)
Pearl Street | Society of Mind Home Page | Dawson Home Page |