Invited Talks

A New Semantics: Merging Propositional and Distributional Information

Eduard Hovy, Director of the Human Language Technology Group, Information Sciences Institute of the University of Southern California
http://www.isi.edu/~hovy/

Before the 1990s, many NLP systems worked with semantic models of one form or another. Today, few do so; those that address semantics at all focus on isolated phenomena such as word senses, entailments, or frame elements. Despite hundreds of years of study on semantics, theories and representations of semantic content "the actual meaning of the symbols used in semantic propositions" remain impoverished. The traditional extensional and intensional models of semantics are difficult to actually flesh out in practice, and no large-scale models of this kind exist. Recently, NLP researchers have increasingly treated topic-based word distributions (also called 'context vectors', 'topic models', 'topic signatures', 'language models', etc.) as a de facto placeholder for semantics at various levels of granularity. This talk argues for a new kind of semantics that combines traditional symbolic logic-based proposition-style semantics (of the kind used in older NLP) with computation-based statistical word distribution information (what is being called Distributional Semantics in modern NLP). The core resource is a single lexico-semantic "lexicon" that can be used for a variety of tasks. I show a simple way to define such a lexicon and how to use it for various tasks. Combining the two views of semantics opens many fascinating questions that beg study, including the operation of logical operators such as negation and modalities over word(sense) distributions, the nature of ontological facets required to define concepts, and the action of compositionality over statistical concepts.

A, is, I, and, the: How our smallest words reveal the most about who we Are

James W. Pennebaker, Centennial Liberal Arts Professor and Chair of Psychology at the University of Texas at Austin
http://www.psy.utexas.edu/pennebaker

Pronouns, articles, and other forgettable function words are the most frequently used and the least understood of all word categories. Analyses of their usage in natural language demonstrate that function words are closely tied to basic social and psychological processes. In every spoken and written language, function words reveal honesty, depression, age, social status, and the relationships among friends, coworkers, and lovers. These findings have applications in the fields of national security, education, dating, advertizing, and mental health. Oh, and Twitter and Facebook.

Program Co-Chairs

• Srinivas Bangalore, AT&T
• Eric Fosler-Lussier, The Ohio State University
• Ellen Riloff, University of Utah