End-to-end Relation Extraction using Neural Networks and Markov Logic Networks

Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers, pages 818–827, Valencia, Spain, April 3-7, 2017. c⃝2017 Association for Computational Linguistics End-to-end Relation Extraction using Neural Networks and Markov Logic Networks Sachin Pawar1,2, Pushpak Bhattacharyya2, and Girish K. Palshikar1 1TCS Research, Tata Consultancy Services, Pune 2Indian Institute of Technology Bombay, Mumbai {sachin7.p,gk.palshikar}@tcs.com pb@cse.iitb.ac.in Abstract End-to-end relation extraction refers to identifying boundaries of entity mentions, entity types of these mentions and appro- priate semantic relation for each pair of mentions. Traditionally, separate predic- tive models were trained for each of these tasks and were used in a “pipeline” fash- ion where output of one model is fed as input to another. But it was observed that addressing some of these tasks jointly re- sults in better performance. We propose a single, joint neural network based model to carry out all the three tasks of bound- ary identification, entity type classifica- tion and relation type classification. This model is referred to as “All Word Pairs” model (AWP-NN) as it assigns an appro- priate label to each word pair in a given sentence for performing end-to-end rela- tion extraction. We also propose to re- fine output of the AWP-NN model by us- ing inference in Markov Logic Networks (MLN) so that additional domain knowl- edge can be effectively incorporated. We demonstrate effectiveness of our approach by achieving better end-to-end relation ex- traction performance than all 4 previous joint modelling approaches, on the stan- dard dataset of ACE 2004. 1 Introduction The task of relation extraction (RE) deals with identifying whether any pre-defined semantic re- lation holds between a pair of entity mentions in the given sentence. Pure relation extraction tech- niques (Zhou et al., 2005; Jiang and Zhai, 2007; Bunescu and Mooney, 2005; Qian et al., 2008) assume that for a sentence, gold-standard entity mentions (i.e. boundaries as well as types) in it are known. In contrast, end-to-end relation extraction deals with plain sentences without assuming any knowledge of entity mentions in them. The task of end-to-end relation extraction consists of three sub-tasks: i) identifying boundaries of entity men- tions, ii) identifying entity types of these mentions and iii) identifying appropriate semantic relation for each pair of mentions. First two sub-tasks correspond to the Entity Detection and Tracking task defined by the the Automatic Content Extrac- tion (ACE) program (Doddington et al., 2004) and the third sub-task corresponds to the Relation De- tection and Characterization (RDC) task. ACE standard defined 7 entity types1: PER (person), ORG (organization), LOC (location), GPE (geo- political entity), FAC (facility), VEH (vehicle) and WEA (weapon). It also defined 7 coarse level relation types2: EMP-ORG (employment), PER- SOC (personal/social), PHYS (physical), GPE- AFF (GPE affiliation), OTHER-AFF (PER/ORG affiliation), ART (agent-artifact) and DISC (dis- course). Traditionally, the three sub-tasks of end-to-end relation extraction are carried out serially in a “pipeline” fashion. In this case, the errors in any sub-task affect subsequent sub-tasks. Another dis- advantage of this “pipeline” approach is …

Content truncated to 500 words