{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/exploring-graph-structured-passage","title":"Exploring Graph-structured Passage Representation for Multi-hop Reading Comprehension with Graph Neural Networks","arxiv_id":"1809.02040","date":"2018-09-06","proceeding":null,"authors":["Linfeng Song","Zhiguo Wang","Mo Yu","Yue Zhang","Radu Florian","Daniel Gildea"],"abstract":"Multi-hop reading comprehension focuses on one type of factoid question,\nwhere a system needs to properly integrate multiple pieces of evidence to\ncorrectly answer a question. Previous work approximates global evidence with\nlocal coreference information, encoding coreference chains with DAG-styled GRU\nlayers within a gated-attention reader. However, coreference is limited in\nproviding information for rich inference. We introduce a new method for better\nconnecting global evidence, which forms more complex graphs compared to DAGs.\nTo perform evidence integration on our graphs, we investigate two recent graph\nneural networks, namely graph convolutional network (GCN) and graph recurrent\nnetwork (GRN). Experiments on two standard datasets show that richer global\ninformation leads to better answers. Our method performs better than all\npublished results on these datasets.","url_abs":"http://arxiv.org/abs/1809.02040v1","url_pdf":"http://arxiv.org/pdf/1809.02040v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"multi-hop-reading-comprehension","task_name":"Multi-Hop Reading Comprehension"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"reading-comprehension","task_name":"Reading Comprehension"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/question-answering-on-complexquestions","task":"Question Answering","dataset":"COMPLEXQUESTIONS","model":"MHQA","rank_in_archive_order":2,"of":2,"metrics":{"F1":"30.1"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-wikihop","task":"Question Answering","dataset":"WikiHop","model":"MHQA","rank_in_archive_order":6,"of":9,"metrics":{"Test":"65.4"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.02040","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}