{"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/end-to-end-knowledge-routed-relational","title":"End-to-End Knowledge-Routed Relational Dialogue System for Automatic Diagnosis","arxiv_id":"1901.10623","date":"2019-01-30","proceeding":null,"authors":["Lin Xu","Qixian Zhou","Ke Gong","Xiaodan Liang","Jianheng Tang","Liang Lin"],"abstract":"Beyond current conversational chatbots or task-oriented dialogue systems that\nhave attracted increasing attention, we move forward to develop a dialogue\nsystem for automatic medical diagnosis that converses with patients to collect\nadditional symptoms beyond their self-reports and automatically makes a\ndiagnosis. Besides the challenges for conversational dialogue systems (e.g.\ntopic transition coherency and question understanding), automatic medical\ndiagnosis further poses more critical requirements for the dialogue rationality\nin the context of medical knowledge and symptom-disease relations. Existing\ndialogue systems (Madotto, Wu, and Fung 2018; Wei et al. 2018; Li et al. 2017)\nmostly rely on data-driven learning and cannot be able to encode extra expert\nknowledge graph. In this work, we propose an End-to-End Knowledge-routed\nRelational Dialogue System (KR-DS) that seamlessly incorporates rich medical\nknowledge graph into the topic transition in dialogue management, and makes it\ncooperative with natural language understanding and natural language\ngeneration. A novel Knowledge-routed Deep Q-network (KR-DQN) is introduced to\nmanage topic transitions, which integrates a relational refinement branch for\nencoding relations among different symptoms and symptom-disease pairs, and a\nknowledge-routed graph branch for topic decision-making. Extensive experiments\non a public medical dialogue dataset show our KR-DS significantly beats\nstate-of-the-art methods (by more than 8% in diagnosis accuracy). We further\nshow the superiority of our KR-DS on a newly collected medical dialogue system\ndataset, which is more challenging retaining original self-reports and\nconversational data between patients and doctors.","url_abs":"http://arxiv.org/abs/1901.10623v2","url_pdf":"http://arxiv.org/pdf/1901.10623v2.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":"end-to-end-knowledge-routed-relational","repo_url":"https://github.com/HCPLab-SYSU/Medical_DS","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"decision-making","task_name":"Decision Making"},{"task_slug":"dialogue-management","task_name":"Dialogue Management"},{"task_slug":"management","task_name":"Management"},{"task_slug":"medical-diagnosis","task_name":"Medical Diagnosis"},{"task_slug":"natural-language-understanding","task_name":"Natural Language Understanding"},{"task_slug":"task-oriented-dialogue-systems","task_name":"Task-Oriented Dialogue Systems"},{"task_slug":"text-generation","task_name":"Text Generation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1901.10623","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}