作者: Tom Verguts
DOI: 10.1162/JOCN.2007.19.3.409
关键词: Artificial intelligence 、 Artificial neural network 、 Assignment problem 、 Supervised training 、 Connectionism 、 Psychology 、 Time windows 、 Cognition 、 Cognitive science 、 Flutter 、 Pattern recognition
摘要: A task that has been intensively studied at the neural level is f lutter discrimination. I argue discrimination entails a combination of temporal assignment problem and quantity comparison problem, propose network model how these problems are solved. The combines unsupervised one-layer supervised training. part clusters input features (stimulus + time window) categorizes resulting clusters. After training, shows good fit with both behavioral properties. New predictions outlined links other cognitive domains pointed out.