Fwd: PhD Final Oral Defense: Stephen Roller, April 28, 2017 (Friday), 2:00 pm, GDC 3.516

[Stephen Roller]

Hi Flare, NLL,

I'd like to invite everyone to my defense in two weeks. I'll bring German crumble cake and coffee!

Best,

Stephen

---------- Forwarded message ---------
From: Katie Dahm <kd...@cs.utexas.edu>
Date: Fri, Apr 14, 2017 at 14:21
Subject: PhD Final Oral Defense: Stephen Roller, April 28, 2017 (Friday), 2:00 pm, GDC 3.516
To:
Cc: gradoffice <grado...@cs.utexas.edu>

PhD Final Oral Defense: Stephen Roller

Date: April 28, 2pm

Location: GDC 3.516

Committee: Katrin Erk (Advisor), Ray Mooney, Risto Miikkulainen, Sebastian Pado (External)

Proposal title: Identifying Lexical Relationships and Entailments with Distributional Semantics

Abstract:

Many modern efforts in Natural Language Understanding depend on rich and powerful semantic representations of words. Systems for sophisticated logical and textual reasoning often depend heavily on lexical resources to provide critical information about relationships between words, but these lexical resources are expensive to create and maintain, and are never fully comprehensive. Distributional Semantics has long offered methods for automatically inducing meaning representations from large corpora, with little or no annotation efforts. The resulting representations are valuable proxies of semantic similarity, but simply knowing two words are similar cannot tell us their relationship, or whether one entails the other.

In this thesis, we consider how methods from Distributional Semantics may be applied to the difficult task of lexical entailment, where one must predict whether one word implies another. We approach this by showing contributions in areas of hypernymy detection, lexical relationship prediction, lexical substitution, and textual entailment. We propose novel experimental setups, models, analysis, and interpretations, which ultimate provide us with a better understanding of both the nature of lexical entailment, as well as the information available within distributional representations.