{"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/180503352","title":"Learning Heterogeneous Knowledge Base Embeddings for Explainable Recommendation","arxiv_id":"1805.03352","date":"2018-05-09","proceeding":null,"authors":["Qingyao Ai","Vahid Azizi","Xu Chen","Yongfeng Zhang"],"abstract":"Providing model-generated explanations in recommender systems is important to\nuser experience. State-of-the-art recommendation algorithms - especially\ncollaborative filtering (CF)-based approaches with shallow or deep models -\nusually work with various unstructured information sources for recommendation,\nsuch as textual reviews, visual images, and various implicit or explicit\nfeedbacks. Though structured knowledge bases were considered in content-based\napproaches, they have been largely ignored recently due to the research focus\non CF approaches. However, structured knowledge exhibit unique advantages in\npersonalized recommendation systems. When the explicit knowledge about users\nand items is considered for recommendation, the system could provide highly\ncustomized recommendations based on users' historical behaviors and the\nknowledge is helpful for providing informed explanations regarding the\nrecommended items. A great challenge for using knowledge bases for\nrecommendation is how to integrate large-scale structured data, while taking\nadvantage of collaborative filtering for highly accurate performance. Recent\nachievements in knowledge-base embedding (KBE) sheds light on this problem,\nwhich makes it possible to learn user and item representations while preserving\nthe structure of their relationship with external knowledge for explanation. In\nthis work, we propose to explain knowledge-base embeddings for explainable\nrecommendation. Specifically, we propose a knowledge-base representation\nlearning framework to embed heterogeneous entities for recommendation, and\nbased on the embedded knowledge base, a soft matching algorithm is proposed to\ngenerate personalized explanations for the recommended items. Experimental\nresults on real-world e-commerce datasets verified the superior recommendation\nperformance and the explainability power of our approach compared with\nstate-of-the-art baselines.","url_abs":"http://arxiv.org/abs/1805.03352v2","url_pdf":"http://arxiv.org/pdf/1805.03352v2.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":"180503352","repo_url":"https://github.com/LunaBlack/KGAT-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"180503352","repo_url":"https://github.com/fumiyo0607/KGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"180503352","repo_url":"https://github.com/giantonia/KGATv2","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"180503352","repo_url":"https://github.com/melissakou/KG-RecSys","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"180503352","repo_url":"https://github.com/xiangwang1223/knowledge_graph_attention_network","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"collaborative-filtering","task_name":"Collaborative Filtering"},{"task_slug":"explainable-recommendation","task_name":"Explainable Recommendation"},{"task_slug":"link-prediction","task_name":"Link Prediction"},{"task_slug":"recommendation-systems","task_name":"Recommendation Systems"},{"task_slug":"representation-learning","task_name":"Representation Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/link-prediction-on-movielens-25m","task":"Link Prediction","dataset":"MovieLens 25M","model":"CFKG","rank_in_archive_order":4,"of":7,"metrics":{"Hits@10":"0.8152","nDCG@10":"0.5196"},"uses_additional_data":false},{"leaderboard":"/sota/link-prediction-on-yelp","task":"Link Prediction","dataset":"Yelp","model":"CFKG","rank_in_archive_order":3,"of":9,"metrics":{"HR@10":"0.8729","nDCG@10":"0.5826"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}