Temporal Information Extraction for Question Answering Using Syntactic Dependencies in an LSTM-based Architecture

Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 887–896 Copenhagen, Denmark, September 7–11, 2017. c⃝2017 Association for Computational Linguistics Temporal Information Extraction for Question Answering Using Syntactic Dependencies in an LSTM-based Architecture Yuanliang Meng, Anna Rumshisky, Alexey Romanov {ymeng,arum,aromanov}@cs.uml.edu Department of Computer Science University of Massachusetts Lowell Lowell, MA 01854 Abstract In this paper, we propose to use a set of simple, uniform in architecture LSTM- based models to recover different kinds of temporal relations from text. Using the shortest dependency path between en- tities as input, the same architecture is im- plemented to extract intra-sentence, cross- sentence, and document creation time rela- tions. A “double-checking” technique re- verses entity pairs in classification, boost- ing the recall of positive cases and reduc- ing misclassifications between opposite classes. An efficient pruning algorithm resolves conflicts globally. Evaluated on QA-TempEval (SemEval2015 Task 5), our proposed technique outperforms state-of- the-art methods by a large margin. We also conduct intrinsic evaluation and post state- of-the-art results on Timebank-Dense. 1 Introduction Recovering temporal information from text is es- sential to many text processing tasks that require deep language understanding, such as answering questions about the timeline of events or auto- matically producing text summaries. This work presents intermediate results of an effort to build a temporal reasoning framework with contemporary deep learning techniques. Until recently, there has been remarkably few attempts to evaluate temporal information extrac- tion (TemporalIE) methods in context of down- stream applications that require reasoning over the temporal representation. One recent effort to conduct such evaluation was SemEval2015 Task 5, a.k.a. QA-TempEval (Llorens et al., 2015a), which used question answering (QA) as the tar- get application. QA-TempEval evaluated systems producing TimeML (Pustejovsky et al., 2003) an- notation based on how well their output could be used in QA. We believe that application-based evaluation of TemporalIE should eventually com- pletely replace the intrinsic evaluation if we are to make progress, and therefore we evaluated our techniques mainly using QA-TempEval setup. Despite the recent advances produced by multi- layer neural network architectures in a variety of areas, the research community is still struggling to make neural architectures work for linguistic tasks that require long-distance dependencies (such as discourse parsing or coreference resolution). Our goal was to see if a relatively simple architecture with minimal capacity for retaining information was able to incorporate the information required to identify temporal relations in text. Specifically, we use several simple LSTM- based components to recover ordering relations between temporally relevant entities (events and temporal expressions). These components are fairly uniform in their architecture, relying on de- pendency relations recovered with a very small number of mature, widely available processing tools, and require minimal engineering otherwise. To our knowledge, this is the first attempt to apply such simplified techniques to the TemporalIE task, and we demonstrate this streamlined architecture is able to outperform state-of-the-art results on a temporal QA task with a large margin. In order to demonstrate generalizability of our proposed architecture, we also …

Content truncated to 500 words