Abstract
Temporal reasoning remains as an unsolved task for Natural Language Processing (NLP), particularly demonstrated in the clinical domain. The complexity of temporal representation in language is evident as results of the 2016 Clinical TempEval challenge indicate: the current state-of-the-art systems perform well in solving mention-identification tasks of event and time expressions but poorly in temporal relation extraction, showing a gap of around 0.25 point below human performance. We explore to adapt the tree-based LSTM-RNN model proposed by Miwa and Bansal (2016) to temporal relation extraction from clinical text, obtaining a five point improvement over the best 2016 Clinical TempEval system and two points over the state-of-the-art. We deliver a deep analysis of the results and discuss the next step towards human-like temporal reasoning.
Highlights
Temporal Information Extraction (TIE) is an active research area in Natural Language Processing (NLP), where the ultimate goal is to be able to represent the development of a story over time
Regardless of the Proceedings of the 9th International Workshop on Health Text Mining and Information Analysis (LOUHI 2018), pages 55–64 Brussels, Belgium, October 31, 2018. c 2018 Association for Computational Linguistics increase in annotation agreement of temporal relations by relying on narrative containers, there is a consensus within the research community regarding TIE difficulty
Clinical language processing represents a special challenge to NLP systems
Summary
Temporal Information Extraction (TIE) is an active research area in NLP, where the ultimate goal is to be able to represent the development of a story over time. TIE is a key to text processing tasks including Question Answering and Text Summarization and follows the traditional pipeline of named entity recognition (NER) and relation extraction separately. Research on this area has been led by TempEval shared tasks (Verhagen et al, 2007, 2010; UzZaman et al, 2013) but in recent years, the target domain has been shifted to the clinical domain. Narrative containers can be thought of as temporal buckets in which an event or series of events may fall. The only corpus annotated with narrative containers is limited to clinical texts
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
