Multiple alternative sentence compressions and word-pair antonymy for automatic text summarization and recognizing textual entailment

TitleMultiple alternative sentence compressions and word-pair antonymy for automatic text summarization and recognizing textual entailment
Publication TypeJournal Articles
Year of Publication2008
AuthorsMohammad S, Dorr BJ, Egan M, Jimmy Lin, Zajic D
JournalProceedings of the Text Analysis Conference (TAC-2008), Gaithersburg, MD
Date Published2008///
Abstract

The University of Maryland participatedin three tasks organized by the Text Anal-
ysis Conference 2008 (TAC 2008): (1)
the update task of text summarization;
(2) the opinion task of text summariza-
tion; and (3) recognizing textual entail-
ment (RTE). At the heart of our summa-
rization system is Trimmer, which gener-
ates multiple alternative compressed ver-
sions of the source sentences that act as
candidate sentences for inclusion in the
summary. For the first time, we investi-
gated the use of automatically generated
antonym pairs for both text summariza-
tion and recognizing textual entailment.
We used an antonymy feature in both the
opinion summarization task and for rec-
ognizing textual entailment. More coher-
ent summaries resulted when using the
antonymy feature as compared to when
not using it. However, performance on
ROUGE dropped. The RTE system per-
formed almost equally well when using
antonyms from WordNet and when using
automatically generated antonyms.