{"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/image-embodied-knowledge-representation","title":"Image-embodied Knowledge Representation Learning","arxiv_id":"1609.07028","date":"2016-09-22","proceeding":null,"authors":["Ruobing Xie","Zhiyuan Liu","Huanbo Luan","Maosong Sun"],"abstract":"Entity images could provide significant visual information for knowledge\nrepresentation learning. Most conventional methods learn knowledge\nrepresentations merely from structured triples, ignoring rich visual\ninformation extracted from entity images. In this paper, we propose a novel\nImage-embodied Knowledge Representation Learning model (IKRL), where knowledge\nrepresentations are learned with both triple facts and images. More\nspecifically, we first construct representations for all images of an entity\nwith a neural image encoder. These image representations are then integrated\ninto an aggregated image-based representation via an attention-based method. We\nevaluate our IKRL models on knowledge graph completion and triple\nclassification. Experimental results demonstrate that our models outperform all\nbaselines on both tasks, which indicates the significance of visual information\nfor knowledge representations and the capability of our models in learning\nknowledge representations with images.","url_abs":"http://arxiv.org/abs/1609.07028v2","url_pdf":"http://arxiv.org/pdf/1609.07028v2.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":"image-embodied-knowledge-representation","repo_url":"https://github.com/thunlp/IKRL","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"knowledge-graph-completion","task_name":"Knowledge Graph Completion"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"triple-classification","task_name":"Triple Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1609.07028","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}