A Search-based Neural Model for Biomedical Nested and Overlapping Event Detection

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 3679–3686, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 3679 A Search-based Neural Model for Biomedical Nested and Overlapping Event Detection Kurt Espinosa1,4, Makoto Miwa2,3, Sophia Ananiadou1 1National Centre for Text Mining, School of Computer Science, The University of Manchester, UK 2Toyota Technological Institute, Nagoya, 468-8511, Japan 3Artificial Intelligence Research Center (AIRC), National Institute of Advanced Industrial Science and Technology (AIST), Japan 4Department of Computer Science, University of the Philippines Cebu, Cebu City, Philippines {kurtjunshean.espinosa, sophia.ananiadou}@manchester.ac.uk makoto-miwa@toyota-ti.ac.jp Abstract We tackle the nested and overlapping event de- tection task and propose a novel search-based neural network (SBNN) structured prediction model that treats the task as a search prob- lem on a relation graph of trigger-argument structures. Unlike existing structured predic- tion tasks such as dependency parsing, the task targets to detect DAG structures, which con- stitute events, from the relation graph. We define actions to construct events and use all the beams in a beam search to detect all event structures that may be overlapping and nested. The search process constructs events in a bottom-up manner while modelling the global properties for nested and overlapping structures simultaneously using neural net- works. We show that the model achieves performance comparable to the state-of-the-art model Turku Event Extraction System (TEES) on the BioNLP Cancer Genetics (CG) Shared Task 2013 without the use of any syntactic and hand-engineered features. Further analyses on the development set show that our model is more computationally efficient while yielding higher F1-score performance. 1 Introduction Nested and overlapping event structures, which occur widely in text, are important because they can capture relations between events such as causality, e.g., a “production” event is a conse- quence of a “discovery” event, which in turn is a result of an “exploration” event. Event extraction involves the identification of a trigger and a set of its arguments in a given text. Figure 1 shows an example of a nested and overlapping event struc- ture in the biomedical domain. The relation graph (topmost) forms a directed acyclic graph (DAG) structure (McClosky et al., 2011) and it encapsu- lates 15 event structures. It contains nested event cause induction mechanisms linking carcinogenesis theme cause induction E1 VEGF cause theme induction E1 Bcl-2 cause theme angiogenesis tumor atloc role influencing induction angiogenesis tumor Bcl-2 VEGF cause theme atloc antagonism Bcl-2 p53 Brn-3a cause theme theme themetheme E1 E2 E3 entity trigger event theme Figure 1: Top: A DAG-structured relation graph (top- most) from the sentence “Looking for mechanisms link- ing Brn-3a to carcinogenesis, we discuss the role of this transcription factor in influencing Bcl-2/VEGF induc- tion of tumor angiogenesis, ...” from BioNLP’13 CG Shared Task (Pyysalo et al., 2015). Bottom: A pair of overlapping and nested events (E2, E3) extracted from the graph with their shared argument event, a flat event (E1). structures such as E2,E3 because one of their ar- …

Content truncated to 500 words