Invited Talk

Abstract

Various people have observed that larges areas of Natural Language Processing have grown further and further apart from the concerns of researchers whose main interest is in language and communication. In this talk I will explore the question of what an optimal collaboration between linguists and computer scientists might look like, in light of the research questions, methods, and tools that both sets of researchers can now offer. The focus of my exploration will be the notion of quantification, which has long been studied by linguists and logicians, but which has not often been the focus of work in modern NLP.

Bio

Kees van Deemter leads the Natural Language Processing group at Utrecht University’s Computing and Information Sciences department. He has led a number of projects in Natural Language Generation, both in industry (at Philips Research from 1984 till 1996) and in academia (at the University of Brighton and at the University of Aberdeen). Working with linguists, psychologists and logicians, he has published widely on theoretical issues in his field, including the monographs Not Exactly: in Praise of Vagueness (Oxford University Press 2010), and Computational Models of Referring: a Study in Cognitive Science (MIT Press 2016).