Forming global representations with extended backpropagation

摘要: The authors present an alternative to fixed microfeature encoding. Meaningful global representations are developed automatically while learning the processing task. When backward error propagation is extended input layer of items evolve reflect underlying relations relevant No microfeatures and no discrete categorization can be seen in resulting representation, i.e., all aspects a concept distributed over whole set units as activity profile. representation determined by contexts where has been encountered, consequently it also these contexts. >