{"pk":31715,"title":"The Learning  of Weak Noun Declension in German: Children vs. Artificial Network Models","subtitle":null,"abstract":"Different artificial networks are presented with\nthe task of learning weak noim declension in\nG e r m a n . This morphological rule is difficult for\ncue-based models because it requires the\nresolution of conflicting cue-predictions and a\ndynamic positional coding due to suffixation. In\naddition to that its 'task frequency* is very low in\nnatural language. This property is preserved in\nthe training input to study the models' abilities to\nhandle low frequency niles. The performances of\nthree kinds of networks:\n1) feedforward networks\n2) recurrent networks\n3) recurrent networks with short term memory\n( S T M ) capacity\nare compared to empirical findings of an\nelicitation experiment with 129 subjects of ages\n5-9 and adult age.","language":"eng","license":{"name":"","short_name":"","text":null,"url":""},"keywords":[],"section":"Submitted Presentations","is_remote":true,"remote_url":"https://escholarship.org/uc/item/3t03s5k3","frozenauthors":[{"first_name":"Peter","middle_name":"","last_name":"Indefrey","name_suffix":"","institution":"Max-Planck-Institute for Psycholinguistics","department":""},{"first_name":"Rainer","middle_name":"","last_name":"Goebel","name_suffix":"","institution":"University of Braunschweig","department":""}],"date_submitted":null,"date_accepted":null,"date_published":"1993-01-01T18:00:00Z","render_galley":null,"galleys":[{"label":"PDF","type":"pdf","path":"https://journalpub.escholarship.org/cognitivesciencesociety/article/31715/galley/22783/download/"}]}