Language and Computation Courses

Advanced Courses

Probabilistic models of world knowledge for language understanding

MH Tessler

Week 2, August 10-14, 2020

This course is intended for students who take Probabilistic Language Understanding in Week 1 who wish to deeped their knowlege and skills of probabilistic models of language understanding.

Abstract:

Structured, probabilistic models of language understanding are booming. These models (e.g., the Rational Speech Act modeling framework) treat language understanding as a probabilistic social reasoning process, wherein speaker and listener jointly coordinate on intended meanings. Crucial to these models is the prior distribution over intended meanings, often considered to be a function of world knowledge and which has been the focus of extensive interest in computational cognitive science. In this course, we will introduce techniques for modeling the rich structure of human concepts and everyday reasoning, which will naturally serve as the background knowledge against which language understanding can faithfully occur. We will review techniques from probabilistic models of cognition and discuss their usage in formal models of pragmatics. The course will focus on the practical implementation of models in a probabilistic programming language (WebPPL; webppl.org) and testing them against intuitions.

Students are strongly encouraged to take Probabilistic Language Understanding in Week 1, as much of the material we cover will build directly on techniques introduced that course. If you do not take that course, you are strongly encouraged be familiar with the basics of programming in WebPPL (e.g., via this short introduction, or dippl.org) as well as with computational models of cognition broadly, and probabilistic or Bayesian models of cognition specifically (e.g., as presented in probmods.org). 

We will use examples from the problang.org and probmods.org web-books.

 

Connect with us

32nd European Summer School in Logic, Language and Information - ESSLLI 2021
Please send any queries to

  • esslli@uu.nl

Search