Introductory Course at ESSLLI 2007
19th European Summer School in Logic, Language and Information
Dublin, Ireland
August 6-17, 2007

 

Introduction to Corpus-based Computational Semantics

 

Alexander Koller Sabine Schulte im Walde
Computer Science Institut für Maschinelle Sprachverarbeitung
Columbia University Universität Stuttgart


Course Description

This course provides an overview of resources and corpus-based approaches in computational lexical semantics. We introduce theoretical issues (such as polysemy), describe related resources (such as WordNet), and present a selection of computational approaches for lexical semantic acquisition (such as learning noun classes from unannotated corpora). We aim for covering the breadth of the field rather than going into detail. Therefore the course presupposes only elementary knowledge of linguistics and machine learning methods.

The course is relevant for linguists to learn about an alternative perspective on semantic information; (computational) logicians to learn about potential ways to obtain world knowledge; and computational linguists to learn how to obtain resources for building better applications.

The topics of the course include:

  1. word senses, polysemy, Wordnet, Word Sense Disambiguation, Senseval
  2. noun similarity/relations/classes, OntoBank
  3. verb similarity/relations/classes, PropBank
  4. selectional preferences, semantic roles, semantic parsing, FrameNet
  5. learning paraphrases and world knowledge, RTE

Course Material