{"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/multi-modal-siamese-network-for-entity","title":"Multi-modal Siamese Network for Entity Alignment","arxiv_id":null,"date":"2022-08-14","proceeding":"KDD 2022 8","authors":["Liyi Chen","Zhi Li","Tong Xu","Han Wu","Zhefeng Wang","Nicholas Jing Yuan","Enhong Chen"],"abstract":"The booming of multi-modal knowledge graphs (MMKGs) has raised the imperative demand for multi-modal entity alignment techniques, which facilitate the integration of multiple MMKGs from separate data sources. Unfortunately, prior arts harness multi-modal knowledge only via the heuristic merging of uni-modal feature embeddings. Therefore, inter-modal cues concealed in multi-modal knowledge could be largely ignored. To deal with that problem, in this paper, we propose a novel Multi-modal Siamese Network for Entity Alignment (MSNEA) to align entities in different MMKGs, in which multi-modal knowledge could be comprehensively leveraged by the exploitation of inter-modal effect. Specifically, we first devise a multi-modal knowledge embedding module to extract visual, relational, and attribute features of entities to generate holistic entity representations for distinct MMKGs. During this procedure, we employ inter-modal enhancement mechanisms to integrate visual features to guide relational feature learning and adaptively assign attention weights to capture valuable attributes for alignment. Afterwards, we design a multi-modal contrastive learning module to achieve inter-modal enhancement fusion with avoiding the overwhelming impact of weak modalities. Experimental results on two public datasets demonstrate that our proposed MSNEA provides state-of-the-art performance with a large margin compared with competitive baselines.","url_abs":"https://dl.acm.org/doi/10.1145/3534678.3539244","url_pdf":"https://dl.acm.org/doi/10.1145/3534678.3539244","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":"multi-modal-siamese-network-for-entity","repo_url":"https://github.com/liyichen-cly/MSNEA","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"contrastive-learning","task_name":"Contrastive Learning"},{"task_slug":"entity-alignment","task_name":"Entity Alignment"},{"task_slug":"knowledge-graphs","task_name":"Knowledge Graphs"},{"task_slug":"multi-modal-entity-alignment","task_name":"Multi-modal Entity Alignment"},{"task_slug":"multi-modal-knowledge-graph","task_name":"Multi-modal Knowledge Graph"}],"methods":[{"method_slug":"align","method_name":"ALIGN"},{"method_slug":"contrastive-learning","method_name":"Contrastive Learning"},{"method_slug":"siamese-network","method_name":"Siamese Network"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-dbp-fr","task":"Multi-modal Entity Alignment","dataset":"UMVM-dbp-fr-en","model":"MSNEA (w/o surf)","rank_in_archive_order":9,"of":10,"metrics":{"Hits@1":"0.583"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-dbp-fr","task":"Multi-modal Entity Alignment","dataset":"UMVM-dbp-fr-en","model":"MSNEA (w/o surf & w/o iter)","rank_in_archive_order":10,"of":10,"metrics":{"Hits@1":"0.557"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-dbp-ja","task":"Multi-modal Entity Alignment","dataset":"UMVM-dbp-ja-en","model":"MSNEA (w/o surf)","rank_in_archive_order":9,"of":10,"metrics":{"Hits@1":"0.557"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-dbp-ja","task":"Multi-modal Entity Alignment","dataset":"UMVM-dbp-ja-en","model":"MSNEA (w/o surf & w/o iter)","rank_in_archive_order":10,"of":10,"metrics":{"Hits@1":"0.541"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-dbp-zh","task":"Multi-modal Entity Alignment","dataset":"UMVM-dbp-zh-en","model":"MSNEA (w/o surf)","rank_in_archive_order":9,"of":10,"metrics":{"Hits@1":"0.648"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-dbp-zh","task":"Multi-modal Entity Alignment","dataset":"UMVM-dbp-zh-en","model":"MSNEA (w/o surf & w/o iter)","rank_in_archive_order":10,"of":10,"metrics":{"Hits@1":"0.609"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-oea-d-w","task":"Multi-modal Entity Alignment","dataset":"UMVM-oea-d-w-v1","model":"MSNEA (w/o surf)","rank_in_archive_order":7,"of":8,"metrics":{"Hits@1":"0.809"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-oea-d-w","task":"Multi-modal Entity Alignment","dataset":"UMVM-oea-d-w-v1","model":"MSNEA (w/o surf & w/o iter)","rank_in_archive_order":8,"of":8,"metrics":{"Hits@1":"0.800"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-oea-d-w-1","task":"Multi-modal Entity Alignment","dataset":"UMVM-oea-d-w-v2","model":"MSNEA (w/o surf)","rank_in_archive_order":7,"of":7,"metrics":{"Hits@1":"0.862"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-oea-en-1","task":"Multi-modal Entity Alignment","dataset":"UMVM-oea-en-de","model":"MSNEA (w/o surf)","rank_in_archive_order":7,"of":8,"metrics":{"Hits@1":"0.788"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-oea-en-1","task":"Multi-modal Entity Alignment","dataset":"UMVM-oea-en-de","model":"MSNEA (w/o surf & w/o iter)","rank_in_archive_order":8,"of":8,"metrics":{"Hits@1":"0.753"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-oea-en","task":"Multi-modal Entity Alignment","dataset":"UMVM-oea-en-fr","model":"MSNEA (w/o surf)","rank_in_archive_order":7,"of":8,"metrics":{"Hits@1":"0.699"},"uses_additional_data":true},{"leaderboard":"/sota/multi-modal-entity-alignment-on-umvm-oea-en","task":"Multi-modal Entity Alignment","dataset":"UMVM-oea-en-fr","model":"MSNEA (w/o surf & w/o iter)","rank_in_archive_order":8,"of":8,"metrics":{"Hits@1":"0.692"},"uses_additional_data":true}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}