TextFlow: A Text Similarity Measure based on Continuous Sequences

Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 763–772 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1071 Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 763–772 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1071 TextFlow: A Text Similarity Measure based on Continuous Sequences Yassine Mrabet mrabety@mail.nih.gov Halil Kilicoglu kilicogluh@mail.nih.gov Lister Hill National Center for Biomedical Communications U.S. National Library of Medicine 8600 Rockville Pike, 20894, Bethesda, MD, USA Dina Demner-Fushman ddemner@mail.nih.gov Abstract Text similarity measures are used in multi- ple tasks such as plagiarism detection, in- formation ranking and recognition of para- phrases and textual entailment. While re- cent advances in deep learning highlighted further the relevance of sequential mod- els in natural language generation, existing similarity measures do not fully exploit the sequential nature of language. Exam- ples of such similarity measures include n- grams and skip-grams overlap which rely on distinct slices of the input texts. In this paper we present a novel text sim- ilarity measure inspired from a common representation in DNA sequence align- ment algorithms. The new measure, called TextFlow, represents input text pairs as continuous curves and uses both the actual position of the words and sequence match- ing to compute the similarity value. Our experiments on eight different datasets show very encouraging results in para- phrase detection, textual entailment recog- nition and ranking relevance. 1 Background The number of pages required to print the content of the World Wide Web was estimated to 305 bil- lion in a 2015 article1. While a big part of this content consists of visual information such as pic- tures and videos, texts also continue growing at a very high pace. A recent study shows that the av- erage webpage weights 1,200 KB with plain text accounting for up to 16% of that size2. While efficient distribution of textual data and computations are the key to deal with the increas- 1http://goo.gl/p9lt7V 2http://goo.gl/c41wpa ing scale of textual search, similarity measures still play an important role in refining search re- sults to more specific needs such as the recognition of paraphrases and textual entailment, plagiarism detection and fine-grained ranking of information. These tasks are also often performed on dedicated document collections for domain-specific applica- tions where text similarity measures can be di- rectly applied. Finding relevant approaches to compute text similarity motivated a lot of research in the last decades (Sahami and Heilman, 2006; Hatzivas- siloglou et al., 1999), and more recently with deep learning methods (Socher et al., 2011; Yih et al., 2011; Severyn and Moschitti, 2015). However, most of the recent advances focused on designing high performance classification methods, trained and tested for specific tasks and did not offer a standalone similarity measure that could be ap- plied (i) regardless of the application domain and (ii) without requiring training corpora. For instance, Yih and Meek (2007) presented an approach to improve text similarity measures for web …

Content truncated to 500 words