{"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/embedding-entities-and-relations-for-learning","title":"Embedding Entities and Relations for Learning and Inference in Knowledge Bases","arxiv_id":"1412.6575","date":"2014-12-20","proceeding":null,"authors":["Bishan Yang","Wen-tau Yih","Xiaodong He","Jianfeng Gao","Li Deng"],"abstract":"We consider learning representations of entities and relations in KBs using\nthe neural-embedding approach. We show that most existing models, including NTN\n(Socher et al., 2013) and TransE (Bordes et al., 2013b), can be generalized\nunder a unified learning framework, where entities are low-dimensional vectors\nlearned from a neural network and relations are bilinear and/or linear mapping\nfunctions. Under this framework, we compare a variety of embedding models on\nthe link prediction task. We show that a simple bilinear formulation achieves\nnew state-of-the-art results for the task (achieving a top-10 accuracy of 73.2%\nvs. 54.7% by TransE on Freebase). Furthermore, we introduce a novel approach\nthat utilizes the learned relation embeddings to mine logical rules such as\n\"BornInCity(a,b) and CityInCountry(b,c) => Nationality(a,c)\". We find that\nembeddings learned from the bilinear objective are particularly good at\ncapturing relational semantics and that the composition of relations is\ncharacterized by matrix multiplication. More interestingly, we demonstrate that\nour embedding-based rule extraction approach successfully outperforms a\nstate-of-the-art confidence-based rule mining approach in mining Horn rules\nthat involve compositional reasoning.","url_abs":"http://arxiv.org/abs/1412.6575v4","url_pdf":"http://arxiv.org/pdf/1412.6575v4.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":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/Sujit-O/pykg2vec","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/awslabs/dgl-ke","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/bi-graph/emgraph","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/facebookresearch/PyTorch-BigGraph","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/giuseppefutia/semi","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/sntcristian/and-kge","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/thiviyant/intelligraphs","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/thu-keg/eakit","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/Accenture/AmpliGraph","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"embedding-entities-and-relations-for-learning","repo_url":"https://github.com/pyg-team/pytorch_geometric/blob/master/torch_geometric/nn/kge/distmult.py","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"link-prediction","task_name":"Link Prediction"}],"methods":[{"method_slug":"transe","method_name":"TransE"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/link-prediction-on-fb15k-237","task":"Link Prediction","dataset":"FB15k-237","model":"DistMult","rank_in_archive_order":73,"of":75,"metrics":{"Hits@10":"0.419","MRR":"0.241"},"uses_additional_data":false},{"leaderboard":"/sota/link-prediction-on-umls","task":"Link Prediction","dataset":"UMLS","model":"DistMult","rank_in_archive_order":10,"of":10,"metrics":{"Hits@10":"0.846","MR":"5.52"},"uses_additional_data":false},{"leaderboard":"/sota/link-prediction-on-wn18","task":"Link Prediction","dataset":"WN18","model":"DistMult","rank_in_archive_order":29,"of":37,"metrics":{"Hits@1":"0.728","Hits@10":"0.936","Hits@3":"0.914","MR":"902","MRR":"0.822"},"uses_additional_data":false},{"leaderboard":"/sota/link-prediction-on-wn18rr","task":"Link Prediction","dataset":"WN18RR","model":"DisMult","rank_in_archive_order":74,"of":75,"metrics":{"Hits@1":"0.39","MRR":"0.43"},"uses_additional_data":false},{"leaderboard":"/sota/link-property-prediction-on-ogbl-biokg","task":"Link Property Prediction","dataset":"ogbl-biokg","model":"DistMult","rank_in_archive_order":14,"of":16,"metrics":{"Ext. data":"No","Number of params":"187648000","Test MRR":"0.8043 ± 0.0003","Validation MRR":"0.8055 ± 0.0003"},"uses_additional_data":false},{"leaderboard":"/sota/link-property-prediction-on-ogbl-wikikg2","task":"Link Property Prediction","dataset":"ogbl-wikikg2","model":"DistMult (500dim)","rank_in_archive_order":27,"of":30,"metrics":{"Ext. data":"No","Number of params":"1250569500","Test MRR":"0.3729 ± 0.0045","Validation MRR":"0.3506 ± 0.0042"},"uses_additional_data":false},{"leaderboard":"/sota/link-property-prediction-on-ogbl-wikikg2","task":"Link Property Prediction","dataset":"ogbl-wikikg2","model":"DistMult (100dim)","rank_in_archive_order":28,"of":30,"metrics":{"Ext. data":"No","Number of params":"250113900","Test MRR":"0.3447 ± 0.0082","Validation MRR":"0.3150 ± 0.0088"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1412.6575","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}