{"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/building-sequential-inference-models-for-end","title":"Building Sequential Inference Models for End-to-End Response Selection","arxiv_id":"1812.00686","date":"2018-12-03","proceeding":null,"authors":["Jia-Chen Gu","Zhen-Hua Ling","Yu-Ping Ruan","Quan Liu"],"abstract":"This paper presents an end-to-end response selection model for Track 1 of the\n7th Dialogue System Technology Challenges (DSTC7). This task focuses on\nselecting the correct next utterance from a set of candidates given a partial\nconversation. We propose an end-to-end neural network based on enhanced\nsequential inference model (ESIM) for this task. Our proposed model differs\nfrom the original ESIM model in the following four aspects. First, a new word\nrepresentation method which combines the general pre-trained word embeddings\nwith those estimated on the task-specific training set is adopted in order to\naddress the challenge of out-of-vocabulary (OOV) words. Second, an attentive\nhierarchical recurrent encoder (AHRE) is designed which is capable to encode\nsentences hierarchically and generate more descriptive representations by\naggregation. Third, a new pooling method which combines multi-dimensional\npooling and last-state pooling is used instead of the simple combination of max\npooling and average pooling in the original ESIM. Last, a modification layer is\nadded before the softmax layer to emphasize the importance of the last\nutterance in the context for response selection. In the released evaluation\nresults of DSTC7, our proposed method ranked second on the Ubuntu dataset and\nthird on the Advising dataset in subtask 1 of Track 1.","url_abs":"http://arxiv.org/abs/1812.00686v1","url_pdf":"http://arxiv.org/pdf/1812.00686v1.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":"building-sequential-inference-models-for-end","repo_url":"https://github.com/JasonForJoy/DSTC7-ResponseSelection","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"conversational-response-selection","task_name":"Conversational Response Selection"},{"task_slug":"descriptive","task_name":"Descriptive"},{"task_slug":"word-embeddings","task_name":"Word Embeddings"}],"methods":[{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"esim","method_name":"ESIM"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/conversational-response-selection-on-dstc7","task":"Conversational Response Selection","dataset":"DSTC7 Ubuntu","model":"Sequential Inference Models","rank_in_archive_order":5,"of":5,"metrics":{"1-of-100 Accuracy":"60.8%"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}