[Colloq] Thesis Proposal: Nugget-based Matching and Evaluation for Information Retrieval and Summarization, November 5th, 1:30 - 2:30

[Fong, Andy]

Wednesday, November 5th, 1:30-2:30pm, 366 WVH

Title:
Nugget-based Matching and Evaluation for Information Retrieval and Summarization

Abstract:
Test collections are a vital part of Information Retrieval research used to
train learning algorithms, evaluate system performance, and create standards
for comparisons across systems. For summarization systems in particular,
evaluation requires a gold standard set of information and a way to match that
information to system-provided summaries. However, current manual methods are
time intensive and automatic methods are naive and inflexible.

I propose to improve these methods by addressing three aspects of the
procedure: 1) the ways in which test collections are constructed to provide
greater information content and reusability with less effort; 2) methods for
matching ideal text to candidate text and evaluating various Information
Retrieval systems, from traditional ad-hoc ranked retrieval to multi-document
summarization systems; 3) methods for accounting for multiple facets of
relevance and additions of variables such as timeliness and verbosity of
information


Matthew Ekstrand-Abueg

Department of Computer Science, 472 WVH

College of Computer and Information Science

Northeastern University


Andrew W. Fong
Program Assistant

Northeastern University
College of Computer and Information Science
360 Huntington Avenue
202 West Village H
Boston, MA 02115
617-373-8493
a.fong at neu.edu

Follow us on Twitter - @CCISGrad<https://twitter.com/CCISGrad>
Like us on Facebook - CCIS Graduate School<https://www.facebook.com/CCISGradSchool?ref_type=bookmark>