Demonstrating systematic behaviour in language processing, i.e. the ability to generalize from a small training set, remains an open challenge for connectionist models as potential candidates for explaining human cognition. In the talk, I will refer to this issue in the context of the syntactic systematicity (deciding whether a novel word sequence is grammatical or not) and the semantic systematicity (generating a semantic representation to the novel sentence). I will discuss the pros and limitations of the proposed recursive neural models with self-organization.
Last modified: Tuesday, 02-Sep-2008 09:04:45 NZST
This page is maintained by the seminar list administrator.