{"pk":33079,"title":"A Connectionist Model for Classification Learning - The lAK Model","subtitle":null,"abstract":"The connectionist model lAK (Information evaluation using \nconfigurations) for classification learning is presented here. \nThe model can be placed between feature based (e.g. Gluck \n& Bower, 1988) and exemplar based models (e.g. ALCOVE , \nKruschke, 1992). Specific to this model is that during \nlearning, sets of input features are probabilistically sampled. \nThese sets are represented, in a hidden layer, by \nconfiguration nodes. These configuration nodes are \nconnected to output nodes that represent category labels. A \nfurther characteristic of the lAK model is a mechanism \nwhich enhances retrieval of information. Simulations with \nthe lAK model can explain different phenomena of \nclassification learning which have been found in \nexperimental studies: A Type 2 advantage without \ndimensional attention learning observed by Shepard et al. \n(1961); a generalisation of prototypes; a generalization based \non similarity to learned exemplars; a differential forgetting \nof prototypes and exemplars; a moderate interference (fan \neffect) caused by stimulus similarity; and the missing of \ncatastrophic interference even in A-B/A-Brdesigns.","language":"eng","license":{"name":"","short_name":"","text":null,"url":""},"keywords":[],"section":"17","is_remote":true,"remote_url":"https://escholarship.org/uc/item/3cd229cq","frozenauthors":[{"first_name":"Martin","middle_name":"","last_name":"Heydemann","name_suffix":"","institution":"Institut fuer Psychologic Technische Hochschule Darmstadt","department":""}],"date_submitted":null,"date_accepted":null,"date_published":"1995-01-01T18:00:00Z","render_galley":null,"galleys":[{"label":"PDF","type":"pdf","path":"https://journalpub.escholarship.org/cognitivesciencesociety/article/33079/galley/24140/download/"}]}