Title: Pragmatic reasoning in large-scale NLP systems

Presenter: Director, Center for the Study of Language and Information
Professor, Linguistics and Computer Science, Stanford University

Website

Abstract: Iterated response models of pragmatic language use (e.g., the Rational Speech Acts model) have achieved an impressive range of results in linguistics and psychology. These models are also proving to be valuable components in large-scale natural language processing systems. This talk will review recent work on such pragmatic NLP hybrids. I will seek to identify effective strategies for designing and assessing these models, and I will show that they encourage us to reformulate traditional NLP problems as genuine communication tasks that are grounded in specific contexts, thereby making such tasks more realistic and more relevant to research in pragmatics.

Bio: Coming Soon