{"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/structured-triplet-learning-with-pos-tag","title":"Structured Triplet Learning with POS-tag Guided Attention for Visual Question Answering","arxiv_id":"1801.07853","date":"2018-01-24","proceeding":null,"authors":["Zhe Wang","Xiaoyi Liu","Liangjian Chen","Li-Min Wang","Yu Qiao","Xiaohui Xie","Charless Fowlkes"],"abstract":"Visual question answering (VQA) is of significant interest due to its\npotential to be a strong test of image understanding systems and to probe the\nconnection between language and vision. Despite much recent progress, general\nVQA is far from a solved problem. In this paper, we focus on the VQA\nmultiple-choice task, and provide some good practices for designing an\neffective VQA model that can capture language-vision interactions and perform\njoint reasoning. We explore mechanisms of incorporating part-of-speech (POS)\ntag guided attention, convolutional n-grams, triplet attention interactions\nbetween the image, question and candidate answer, and structured learning for\ntriplets based on image-question pairs. We evaluate our models on two popular\ndatasets: Visual7W and VQA Real Multiple Choice. Our final model achieves the\nstate-of-the-art performance of 68.2% on Visual7W, and a very competitive\nperformance of 69.6% on the test-standard split of VQA Real Multiple Choice.","url_abs":"http://arxiv.org/abs/1801.07853v1","url_pdf":"http://arxiv.org/pdf/1801.07853v1.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":"structured-triplet-learning-with-pos-tag","repo_url":"https://github.com/wangzheallen/STL-VQA","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"multiple-choice","task_name":"Multiple-choice"},{"task_slug":"pos","task_name":"POS"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"tag","task_name":"TAG"},{"task_slug":null,"task_name":"Triplet"},{"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":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1801.07853","atlas_url":"https://app.syntology.ai/?focus=1801.07853","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}