Biased LexRank : Passage Retrieval using Random Walks with Question-Based Priors
description
Transcript of Biased LexRank : Passage Retrieval using Random Walks with Question-Based Priors
Intelligent Database Systems Lab
Presenter : JHOU, YU-LIANG
Authors : Jahna Otterbacher a , Gunes Erkan b , Dragomir R. Radev
2009, IPM
Biased LexRank: Passage Retrieval usingRandom Walks with Question-Based Priors
Intelligent Database Systems Lab
Outlines
MotivationObjectivesMethodology Experimental ResultConclusionsComments
Intelligent Database Systems Lab
Motivation• Text summarization is one of the hardest problems in
information retrieval, because it is not very well-defined.
• There are various definitions of text summarization resulting from different approaches to solving the problem.
• There is often no agreement as to what a good summary is even when we are dealing with a particular definition of the problem.
Intelligent Database Systems Lab
ObjectivesUsing biased LexRank on achieving text summarization and retrieval QA more effect .
Intelligent Database Systems Lab
LexRank
Intelligent Database Systems Lab
Biased LexRank
Intelligent Database Systems Lab
Biased LexRank-application-QA
QA systems is to retrieve the sentences that potentially contain the answer to the question .
Intelligent Database Systems Lab
Passage retrieval-summarization
Computing link weights
Intelligent Database Systems Lab
Passage retrieval- Question answering
Intelligent Database Systems Lab
Experimental-result
biased LexRank v.s human summarizers
Intelligent Database Systems Lab
Experimental-QAcorpus
Intelligent Database Systems Lab
Experimental Result effect of similarity for QA
Intelligent Database Systems Lab
Experimental Result LexRank Versus the Baseline Approach
Intelligent Database Systems Lab
Experimental ResultLexRank Versus the Baseline Approach
Intelligent Database Systems Lab
Conclusions
In the paper, we have also demonstrated the
effectiveness of our method as applied to two classical
IR problems, extractive text summarization and passage
retrieval for question answering.
Intelligent Database Systems Lab
CommentsI think the method improved retrieval performance and comparable to human summarizers.Applications
- Text summarization- Information retrieval