{"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/learning-to-disambiguate-by-asking","title":"Learning to Disambiguate by Asking Discriminative Questions","arxiv_id":"1708.02760","date":"2017-08-09","proceeding":"ICCV 2017 10","authors":["Yining Li","Chen Huang","Xiaoou Tang","Chen-Change Loy"],"abstract":"The ability to ask questions is a powerful tool to gather information in\norder to learn about the world and resolve ambiguities. In this paper, we\nexplore a novel problem of generating discriminative questions to help\ndisambiguate visual instances. Our work can be seen as a complement and new\nextension to the rich research studies on image captioning and question\nanswering. We introduce the first large-scale dataset with over 10,000\ncarefully annotated images-question tuples to facilitate benchmarking. In\nparticular, each tuple consists of a pair of images and 4.6 discriminative\nquestions (as positive samples) and 5.9 non-discriminative questions (as\nnegative samples) on average. In addition, we present an effective method for\nvisual discriminative question generation. The method can be trained in a\nweakly supervised manner without discriminative images-question tuples but just\nexisting visual question answering datasets. Promising results are shown\nagainst representative baselines through quantitative evaluations and user\nstudies.","url_abs":"http://arxiv.org/abs/1708.02760v1","url_pdf":"http://arxiv.org/pdf/1708.02760v1.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":"benchmarking","task_name":"Benchmarking"},{"task_slug":"image-captioning","task_name":"Image Captioning"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"question-generation","task_name":"Question Generation"},{"task_slug":"question-generation","task_name":"Question-Generation"},{"task_slug":"visual-question-answering-1","task_name":"Visual Question Answering"},{"task_slug":"visual-question-answering","task_name":"Visual Question Answering (VQA)"}],"methods":[],"datasets_introduced":[{"slug":"vdqg","name":"VDQG","full_name":"Visual Discriminative Question Generation"}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1708.02760","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}