Universiteit van Amsterdam

Events

Institute for Logic, Language and Computation

Please note that this newsitem has been archived, and may contain outdated information or links.

24 May 2019, IWCS’19 Workshop on Vector Semantics for Dialogue and Discourse (VSDD)

Date: Friday 24 May 2019
Location: Gothenburg, Sweden
Target audience: researchers using vector space methods for semantics in the realm of sentence, dialogue, and discourse
Deadline: Sunday 14 April 2019

Vector models of meaning have revolutionised computational linguistics via their ability to reflect semantic similarities and regularities while providing flexibility to model dynamics and change. However, despite much recent interest inextending these models from the level of words to larger phrases and sentences, there has been little attempt to extend these to the realm of discourse and dialogue. This requires a shift in perspective, moving beyond the static word/sentence view of language to a dynamic perspective seeing language as a set of mechanisms for real-time interactions, encompassing a whole range of actions both sub- and supra-sentential.

This workshop hopes to bring together researchers using vector space methods for semantics in the realm of dialogue and discourse to discuss these challenges and fill this gap. We are planning this event as an open discussion session, with three keynote speakers (see below) and otherwise individual participants having a short slot to discuss a presented poster, with plenty of time allowed for free interactive exchange.

Abstract Format: Abstracts should: be anonymous; be no longer than 2 A4 pages including references; be in .pdf format; use single line spacing, 11pt font with 1-inch margins.

Topics of interest: Distributional Semantics, Vector Space Semantics, Composition Functions, Word Embeddings in Dialogue Systems, Distributional Pragmatics, Vector Composition beyond the Sentence, Sub-sentential Vector Composition, Incremental, Semantic Parsing in Vector Space, Underspecification, Hybrid Symbolic-Distributional Models, Representation Learning for Dialogue & Discourse, Vector Space Models of Dialogue & Discourse Context, Language Grounding.

For more information, see https://sites.google.com/site/dialoguevector/ or contact Mehrnoosh Sadrzadeh at .

Please note that this newsitem has been archived, and may contain outdated information or links.