{"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/dynamic-time-aware-attention-to-speaker-roles","title":"Dynamic Time-Aware Attention to Speaker Roles and Contexts for Spoken Language Understanding","arxiv_id":"1710.00165","date":"2017-09-30","proceeding":null,"authors":["Po-Chun Chen","Ta-Chung Chi","Shang-Yu Su","Yun-Nung Chen"],"abstract":"Spoken language understanding (SLU) is an essential component in\nconversational systems. Most SLU component treats each utterance independently,\nand then the following components aggregate the multi-turn information in the\nseparate phases. In order to avoid error propagation and effectively utilize\ncontexts, prior work leveraged history for contextual SLU. However, the\nprevious model only paid attention to the content in history utterances without\nconsidering their temporal information and speaker roles. In the dialogues, the\nmost recent utterances should be more important than the least recent ones.\nFurthermore, users usually pay attention to 1) self history for reasoning and\n2) others' utterances for listening, the speaker of the utterances may provides\ninformative cues to help understanding. Therefore, this paper proposes an\nattention-based network that additionally leverages temporal information and\nspeaker role for better SLU, where the attention to contexts and speaker roles\ncan be automatically learned in an end-to-end manner. The experiments on the\nbenchmark Dialogue State Tracking Challenge 4 (DSTC4) dataset show that the\ntime-aware dynamic role attention networks significantly improve the\nunderstanding performance.","url_abs":"http://arxiv.org/abs/1710.00165v2","url_pdf":"http://arxiv.org/pdf/1710.00165v2.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":"dynamic-time-aware-attention-to-speaker-roles","repo_url":"https://github.com/MiuLab/Time-SLU","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"dialogue-state-tracking","task_name":"Dialogue State Tracking"},{"task_slug":"spoken-language-understanding","task_name":"Spoken Language Understanding"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1710.00165","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}