University of Cambridge > Talks.cam > Language Technology Lab Seminars > Building a True Semantic World: Generalizing set-theoretic semantics in vector spaces

Building a True Semantic World: Generalizing set-theoretic semantics in vector spaces

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Mohammad Taher Pilehvar.

Quantification is intrinsic to most utterances in natural language. It encompasses fundamental information which is a prerequisite for lexical semantics and inference tasks, such as hyponomy and entailment. Yet, we are unable to denote this information computationally in state-of-the-art semantic models. In this talk, I present an approach to automatically map a distributional semantic space onto a set-theoretic model to induce generalized quantifiers for subject-predicate pairs. I begin from the prediction that there is a functional relationship between distributional information and vectorial concept representations. I first introduce a large, computationally-friendly dataset motivated by understanding, theoretically, to which extent speakers agree on a single model of the world. Then I present a model to test this prediction and show that we can, indeed, map between formalisms, and further, we can generate natural language quantifiers sensibly.

This talk is part of the Language Technology Lab Seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2019 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity