Invited Speakers
- Raquel Fernandez, Institute for Logic, Language & Computation, University of Amsterdam
Title: Common Ground and Audience Design in Referential Games
Abstract: In conversation, we decide what to say and how to say it on the basis of what we share with our dialogue partner. Yet, it is an open question how such accommodation can be modelled in computational agents. Taking a visually grounded referential game as test bed, in this talk I will present recent work where we use computational methods to analyse repeated references exhibiting lexical entrainment and model audience-aware adaptation between agents with asymmetric knowledge. - Lucia Donatelli, Assistant Professor at the Computational Linguistics and Text Mining Lab (CLTL) at Vrije Universiteit in Amsterdam
Title: Compositionality and Its Discontents
Abstract: How do we bridge the gap between compositional semantics and broader notions ofmeaning in computational linguistics? In this talk, I will address this question from several angles. First, I will look at the challenge of adequately representing semantic structure when designing meaning representations, given distinct theoretical and practical considerations. I will present a semantic parsing methodology for normalizing discrepancies between representations at the compositional level to better understand which design differences are semantically rooted and which are superficial. Next, I will present work discussing how representting compositional structure helps on tasks such as cross-lingual parsing and compositional generalization. Finally, I will discuss implications of structured representations and models for generalizing to applications such as situated dialogue and interaction, where compositional semantics alone seems insufficient for robust performance.
