An automaton framework for neural nets that learn

摘要: Brindley (1967, 1969, 1972) has discussed nets of several types formal neurons, many whose functions are modifiable by their own input stimuli. Because Brindley's results widely referred to, for example Marr (1970, 1971) and include some the scarce non-trivial theorems on learning nets, it is important that serious side-conditions be made explicit. The language finite automata used to mathematicize problem adaptation sufficiently remove ambiguities approach. We close paper relating our framework other studies adaptation.