Keyphrase Generation for Scientific Document Retrieval

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1118–1126 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 1118 Keyphrase Generation for Scientific Document Retrieval Florian Boudin1, Ygor Gallina1, Akiko Aizawa2 1LS2N, Universit´e de Nantes, France first.last@univ-nantes.fr 2National Institute of Informatics, Tokyo, Japan aizawa@nii.ac.jp Abstract Sequence-to-sequence models have lead to sig- nificant progress in keyphrase generation, but it remains unknown whether they are reli- able enough to be beneficial for document re- trieval. This study provides empirical evi- dence that such models can significantly im- prove retrieval performance, and introduces a new extrinsic evaluation framework that al- lows for a better understanding of the limi- tations of keyphrase generation models. Us- ing this framework, we point out and dis- cuss the difficulties encountered with supple- menting documents with –not present in text– keyphrases, and generalizing models across domains. Our code is available at https:// github.com/boudinfl/ir-using-kg. 1 Introduction With the exponential growth of the scientific liter- ature (Bornmann and Mutz, 2015), retrieving rel- evant scientific papers becomes increasingly dif- ficult. Keywords, also referred to as keyphrases, provide an effective way to supplement paper index- ing and improve retrieval effectiveness in scientific digital libraries (Barker et al., 1972; Zhai, 1997; Gutwin et al., 1999; Lu and Kipp, 2014). However, only few documents have assigned keyphrases, and those who do were, for the most part, self- labeled by their authors, thus exhibiting annota- tion inconsistencies (Strader, 2011; Suzuki et al., 2011). This has motivated an active line of research on automatic keyphrase extraction (see Hasan and Ng (2014) for an overview) and, more recently, keyphrase generation (Meng et al., 2017), where the task is to find a set of words and phrases that represents the main content of a document. Although models for predicting keyphrases have been extensively evaluated on their ability to re- produce author’s keywords, it still remains unclear whether they can be usefully applied in informa- tion retrieval. One reason for this lack of evidence may have been their relatively low performance dis- couraging attempts at using them for indexing (Liu et al., 2010; Hasan and Ng, 2014). Yet, recently proposed models not only achieve much better per- formance, but also display a property that may have a significant impact on retrieval effectiveness: the capacity to generate keyphrases that do not appear in the source text. These absent keyphrases do not just highlight the topics that are most relevant, but provide some form of semantic expansion by adding new content (e.g. synonyms, semantically related terms) to the index (Greulich, 2011). The goal of this paper is two-fold: to gather empirical evidence as to whether current keyphrase genera- tion models are good enough to improve scientific document retrieval, and to gain further insights into the performance of these models from an extrinsic perspective. Our contributions are listed as follows: • We report significant improvements for strong retrieval models on a standard benchmark col- lection, showing that keyphrases produced by state-of-the-art models are consistently help- ful …

Content truncated to 500 words