{"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/response-ranking-with-deep-matching-networks","title":"Response Ranking with Deep Matching Networks and External Knowledge in Information-seeking Conversation Systems","arxiv_id":"1805.00188","date":"2018-05-01","proceeding":null,"authors":["Liu Yang","Minghui Qiu","Chen Qu","Jiafeng Guo","Yongfeng Zhang","W. Bruce Croft","Jun Huang","Haiqing Chen"],"abstract":"Intelligent personal assistant systems with either text-based or voice-based\nconversational interfaces are becoming increasingly popular around the world.\nRetrieval-based conversation models have the advantages of returning fluent and\ninformative responses. Most existing studies in this area are on open domain\n\"chit-chat\" conversations or task / transaction oriented conversations. More\nresearch is needed for information-seeking conversations. There is also a lack\nof modeling external knowledge beyond the dialog utterances among current\nconversational models. In this paper, we propose a learning framework on the\ntop of deep neural matching networks that leverages external knowledge for\nresponse ranking in information-seeking conversation systems. We incorporate\nexternal knowledge into deep neural models with pseudo-relevance feedback and\nQA correspondence knowledge distillation. Extensive experiments with three\ninformation-seeking conversation data sets including both open benchmarks and\ncommercial data show that, our methods outperform various baseline methods\nincluding several deep text matching models and the state-of-the-art method on\nresponse selection in multi-turn conversations. We also perform analysis over\ndifferent response types, model variations and ranking examples. Our models and\nresearch findings provide new insights on how to utilize external knowledge\nwith deep neural models for response selection and have implications for the\ndesign of the next generation of information-seeking conversation systems.","url_abs":"http://arxiv.org/abs/1805.00188v3","url_pdf":"http://arxiv.org/pdf/1805.00188v3.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":[{"paper_slug":"response-ranking-with-deep-matching-networks","repo_url":"https://github.com/yangliuy/NeuralResponseRanking","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"text-matching","task_name":"Text Matching"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1805.00188","atlas_url":"https://app.syntology.ai/?focus=1805.00188","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}