Context-Aware Latent Dirichlet Allocation for Topic Segmentation

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We propose a new generative model for topic segmentation based on Latent Dirichlet Allocation. The task is to divide a document into a sequence of topically coherent segments, while preserving long topic change-points (coherency) and keeping short topic segments from getting merged (saliency). Most of the existing models either fuse topic segments by keywords or focus on modeling word co-occurrence patterns without merging. They can hardly achieve both coherency and saliency since many words have high uncertainties in topic assignments due to their polysemous nature. To solve this problem, we introduce topic-specific co-occurrence of word pairs within contexts in modeling, to generate more coherent segments and alleviate the influence of irrelevant words on topic assignment. We also design an optimization algorithm to eliminate redundant items in the generated topic segments. Experimental results show that our proposal produces significant improvements in both topic coherence and topic segmentation.

Original languageEnglish
Title of host publicationAdvances in Knowledge Discovery and Data Mining - 24th Pacific-Asia Conference, PAKDD 2020, Proceedings
EditorsHady W. Lauw, Ee-Peng Lim, Raymond Chi-Wing Wong, Alexandros Ntoulas, See-Kiong Ng, Sinno Jialin Pan
PublisherSpringer
Pages475-486
Number of pages12
ISBN (Print)9783030474256
DOIs
Publication statusPublished - Jan 1 2020
Event24th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2020 - Singapore, Singapore
Duration: May 11 2020May 14 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12084 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference24th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2020
CountrySingapore
CitySingapore
Period5/11/205/14/20

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Context-Aware Latent Dirichlet Allocation for Topic Segmentation'. Together they form a unique fingerprint.

  • Cite this

    Li, W., Matsukawa, T., Saigo, H., & Suzuki, E. (2020). Context-Aware Latent Dirichlet Allocation for Topic Segmentation. In H. W. Lauw, E-P. Lim, R. C-W. Wong, A. Ntoulas, S-K. Ng, & S. J. Pan (Eds.), Advances in Knowledge Discovery and Data Mining - 24th Pacific-Asia Conference, PAKDD 2020, Proceedings (pp. 475-486). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12084 LNAI). Springer. https://doi.org/10.1007/978-3-030-47426-3_37