Text segmentation: A topic modeling perspective

Misra, Hemant; Yvon, François; Cappé, Olivier; Jose, Joemon M.

doi:10.1016/j.ipm.2010.11.008

Cited by 55 publications

(39 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This suggests an approach based on document passages. Some previous work on topic segmentation in a document [15,16,17] and passage-based retrieval [18,19] has validated this idea. Here we use it in translation model training.…”

Section: Position-aligned Translation Modelmentioning

confidence: 85%

Position-Aligned Translation Model for Citation Recommendation

Nie

Yang

et al. 2012

String Processing and Information Retrieval

View full text Add to dashboard Cite

Abstract. The goal of a citation recommendation system is to suggest some references for a snippet in an article or a book, and this is very useful for both authors and the readers. The citation recommendation problem can be cast as an information retrieval problem, in which the query is the snippet from an article, and the relevant documents are the cited articles. In reality, the citation snippet and the cited articles may be described in different terms, and this makes the citation recommendation task difficult. Translation model is very useful in bridging the vocabulary gap between queries and documents in information retrieval. It can be trained on a collection of query and document pairs, which are assumed to be parallel. However, such training data contains much noise: a relevant document usually contains some relevant parts along with irrelevant ones. In particular, the citation snippet may only mention only some parts of the cited article's content. To cope with this problem, in this paper, we propose a method to train translation models on such noisy data, called position-aligned translation model. This model tries to align the query to the most relevant parts of the document, so that the estimated translation probabilities could rely more on them. We test this model in a citation recommendation task for scientific papers. Our experiments show that the proposed method can significantly improve the previous retrieval methods based on translation models.

show abstract

Section: Position-aligned Translation Modelmentioning

confidence: 85%

Position-Aligned Translation Model for Citation Recommendation

Nie

Yang

et al. 2012

String Processing and Information Retrieval

View full text Add to dashboard Cite

show abstract

“…The algorithm combines the segmentation probability model of Eisenstein with the non uniform prior on segmentations from (Utiyama & Isahara, 2001). Misra et al (Misra, Yvon, Cappe, & Jose, 2011) adopt a similar approach and use a segment prior similar to that of Utiyama, but consider segmentation probabilities based on latent Dirichlet allocation and multinomial mixture models. The Bayesian segmentation algorithm in Section 3 could be replaced with Misra's algorithm.…”

Section: Discussionmentioning

confidence: 99%

An improved algorithm for unsupervised decomposition of a multi‐author document

Giannella

2015

Asso for Info Science & Tech

View full text Add to dashboard Cite

Abstract AbstractThis paper addresses the problem of unsupervised decomposition of a multi author text document: identifying the sentences that were written by each author assuming the number of authors is unknown. An approach, BayesAD, is developed for solving this problem: apply a Bayesian segmentation algorithm, followed by a segment BayesAD exhibited greater accuracy than AK in all experiments. However, BayesAD has a parameter that needs to be set and which had a non trivial impact on accuracy.Developing an effective method for eliminating this need would be a fruitful direction for future work. When controlling for topic, the accuracy of BayesAD and AK were, in all but one case, worse than a baseline approach wherein one author was assumed to write all sentences in the input text document. Hence, room for improved solutions exists.

show abstract

“…More recently, topic modeling (notably LDA) has been applied to discourse segmentation as well (e.g. Misra et al (2011); see also Riedl and Biemann (2012) for an overview). The dominant interest is on topical shifts in text as indicator of discourse structure, however topic modeling estimation is computationally expensive and needs domain-adaptation.…”

Section: Related Workmentioning

confidence: 99%

Modeling intra-textual variation with entropy and surprisal: topical vs. stylistic patterns

Degaetano-Ortlieb

Teich

2017

Proceedings of the Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanit

View full text Add to dashboard Cite

We present a data-driven approach to investigate intra-textual variation by combining entropy and surprisal. With this approach we detect linguistic variation based on phrasal lexico-grammatical patterns across sections of research articles. Entropy is used to detect patterns typical of specific sections. Surprisal is used to differentiate between more and less informationally-loaded patterns as well as types of information (topical vs. stylistic). While we here focus on research articles in biology/genetics, the methodology is especially interesting for digital humanities scholars, as it can be applied to any text type or domain and combined with additional variables (e.g. time, author or social group) to obtain insights on intra-textual variation.

show abstract

Text segmentation: A topic modeling perspective

Cited by 55 publications

References 25 publications

Position-Aligned Translation Model for Citation Recommendation

Position-Aligned Translation Model for Citation Recommendation

An improved algorithm for unsupervised decomposition of a multi‐author document

Modeling intra-textual variation with entropy and surprisal: topical vs. stylistic patterns

Contact Info

Product

Resources

About