Probabilistic models of similarity and plausibility in context

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

• Diarmuid Ó Séaghdha, University of Cambridge
• Friday 02 December 2011, 12:00-13:00
• FW26, Computer Laboratory.

If you have a question about this talk, please contact Ekaterina Kochmar.

The distributional approach in its many guises is the most popular paradigm for current research on lexical semantics. In this talk I’ll describe a framework for distributional semantics based on latent variable probabilistic models of co-occurrence (aka “topic models”). These models can answer a variety of semantic questions about how a word interacts with its context; I will focus on questions about co-occurrence plausibility and about similarity between words in the disambiguating context of a sentence or syntactic structure. Modelling plausibility corresponds to the well-known task of selectional preference learning; in-context similarity is fundamental to disambiguation tasks such as lexical substitution. I will show that relatively simple topic models give very good performance across a range of lexical semantic evaluation settings.

This talk is part of the NLIP Seminar Series series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge Forum of Science and Humanities
• Cambridge Language Sciences
• Cambridge talks
• Chris Davis' list
• Computer Education Research
• Computing Education Research
• Department of Computer Science and Technology talks and seminars
• FW26, Computer Laboratory
• Graduate-Seminars
• Guy Emerson's list
• Interested Talks
• Language Sciences for Graduate Students
• NLIP Seminar Series
• PMRFPS's
• Simon Baker's List
• Trust & Technology Initiative - interesting events
• bld31
• ndk22's list
• ob366-ai4er
• rp587

Note that ex-directory lists are not shown.