BME HLT | Computational Lexical Semantics

Computational Lexical Semantics

Eötvös Loránd University, 2017/2018 spring

Reading seminar at the theoretical linguistics program of the Eötvös Loránd University and the Hungarian Academy of Sciences. Note that the Hungarian version of this page is more informative and up-to-date.

Purpose of course

Most of the content of texts is contained in the meaning of words. The reading seminar will give a glimpse of the huge theoretical tradition and tools gathered since Katz and Fodor (1963), and also tries to give a picture of the present situation.

Content of the course

See also the reading list.

  • lexical decomposition (Katz and Fodor, 1963)
  • the golden years of artificial intelligence research
  • cognitively inspired approaches (Jackendoff, Wierzbicka, Talmy, Pustejovsky)
  • semantics in present natural language processing

Assessment and evaluation

Students will process the papers and present them to the class. The speaker will ask a few homework questions, and evaluate the solutions given by others. Computer experiments in connection to the presented theories with the supervision of the instructor are also encouraged but not required.

If all students understand Hungarian, courses will be in Hungarian, otherwise English.


Symbolic representations
  1. Lexical decomposition (Katz and Fodor, 1963)
  2. Semantic networks, 2 weeks (Quillian, 1969, Schank, 1972, Woods, 1975, Charniak 1972)
  3. Commonsense knowledge: Conceptual Graphs (Sowa, 1976), Cyc (Lenat and Guha, 1990), KL-ONE (Brachman and Levesque, 1985).
  4. the cognitively inspired, 2 weeks (Conceptual Semantics (Jackendoff 1990); Natural Semantic Metalanguage (Wierzbicka, 1972); Force dynamics (Talmy 1988); The generative lexicon (Pustejovsky, 1995))
  5. computational lexicography (Boguraev and Briscoe, 1989 chapters 1 and 9)
  6. The state of the art: WordNet, MEO, ConceptNet, Deep Lexical Semantics, Abstract Meaning Representation, 4lang
Distributional models
  1. Harris (e.g. 1951)
  2. Osgood (1957)
  3. Latent semantic analysis (Landauer and Dumais, 1997; Deerwester et al., 1990; Salton et al., 1975) and mutual information (Church 1990, Levy and Goldberg 2014+)
  4. vector-space language models (VSMs, Sahlgren, 2006; Turney and Pantel, 2010, Cilibrasi 2007)
  5. psychological reality of VSMs (Mitchell et al., 2008)
  6. language modeling (Brown et al 1992, Bengio 2003, Mikolov et al 2013a, b, c, Levy and Goldberg 2014+)