{"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/sc-block-supervised-contrastive-blocking","title":"SC-Block: Supervised Contrastive Blocking within Entity Resolution Pipelines","arxiv_id":"2303.03132","date":"2023-03-06","proceeding":null,"authors":["Alexander Brinkmann","Roee Shraga","Christian Bizer"],"abstract":"The goal of entity resolution is to identify records in multiple datasets that represent the same real-world entity. However, comparing all records across datasets can be computationally intensive, leading to long runtimes. To reduce these runtimes, entity resolution pipelines are constructed of two parts: a blocker that applies a computationally cheap method to select candidate record pairs, and a matcher that afterwards identifies matching pairs from this set using more expensive methods. This paper presents SC-Block, a blocking method that utilizes supervised contrastive learning for positioning records in the embedding space, and nearest neighbour search for candidate set building. We benchmark SC-Block against eight state-of-the-art blocking methods. In order to relate the training time of SC-Block to the reduction of the overall runtime of the entity resolution pipeline, we combine SC-Block with four matching methods into complete pipelines. For measuring the overall runtime, we determine candidate sets with 99.5% pair completeness and pass them to the matcher. The results show that SC-Block is able to create smaller candidate sets and pipelines with SC-Block execute 1.5 to 2 times faster compared to pipelines with other blockers, without sacrificing F1 score. Blockers are often evaluated using relatively small datasets which might lead to runtime effects resulting from a large vocabulary size being overlooked. In order to measure runtimes in a more challenging setting, we introduce a new benchmark dataset that requires large numbers of product offers to be blocked. On this large-scale benchmark dataset, pipelines utilizing SC-Block and the best-performing matcher execute 8 times faster than pipelines utilizing another blocker with the same matcher reducing the runtime from 2.5 hours to 18 minutes, clearly compensating for the 5 minutes required for training SC-Block.","url_abs":"https://arxiv.org/abs/2303.03132v2","url_pdf":"https://arxiv.org/pdf/2303.03132v2.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":"sc-block-supervised-contrastive-blocking","repo_url":"https://github.com/wbsg-uni-mannheim/sc-block","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"blocking","task_name":"Blocking"},{"task_slug":"contrastive-learning","task_name":"Contrastive Learning"},{"task_slug":"entity-resolution","task_name":"Entity Resolution"}],"methods":[{"method_slug":"contrastive-learning","method_name":"Contrastive Learning"}],"datasets_introduced":[{"slug":"wdc-block","name":"WDC Block","full_name":"WDC Block: A Blocking Benchmark"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/blocking-on-abt-buy","task":"Blocking","dataset":"Abt-Buy","model":"SC-Block","rank_in_archive_order":2,"of":6,"metrics":{"Candidate Set Size":"5000","Recall":"99.5"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-abt-buy","task":"Blocking","dataset":"Abt-Buy","model":"BM25","rank_in_archive_order":3,"of":6,"metrics":{"Candidate Set Size":"8000","Recall":"94.7"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-amazon-google","task":"Blocking","dataset":"Amazon-Google","model":"SC-Block","rank_in_archive_order":1,"of":6,"metrics":{"Candidate Set Size":"11000","Recall":"99.6"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-amazon-google","task":"Blocking","dataset":"Amazon-Google","model":"BM25","rank_in_archive_order":3,"of":6,"metrics":{"Candidate Set Size":"40000","Recall":"98.7"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-wdc-block-large","task":"Blocking","dataset":"WDC Block - large","model":"SC-Block","rank_in_archive_order":1,"of":2,"metrics":{"Candidate Set Size":"5000000","Recall":"89.5"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-wdc-block-large","task":"Blocking","dataset":"WDC Block - large","model":"BM25","rank_in_archive_order":2,"of":2,"metrics":{"Candidate Set Size":"20000000","Recall":"95.5"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-wdc-block-medium","task":"Blocking","dataset":"WDC Block - medium","model":"SC-Block","rank_in_archive_order":1,"of":2,"metrics":{"Candidate Set Size":"100000","Recall":"91.9"},"uses_additional_data":true},{"leaderboard":"/sota/blocking-on-wdc-block-medium","task":"Blocking","dataset":"WDC Block - medium","model":"BM25","rank_in_archive_order":2,"of":2,"metrics":{"Candidate Set Size":"500000","Recall":"97.8"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-wdc-block-small","task":"Blocking","dataset":"WDC Block - small","model":"BM25","rank_in_archive_order":1,"of":2,"metrics":{"Candidate Set Size":"250000","Recall":"96.9%"},"uses_additional_data":false},{"leaderboard":"/sota/blocking-on-wdc-block-small","task":"Blocking","dataset":"WDC Block - small","model":"SC-Block","rank_in_archive_order":2,"of":2,"metrics":{"Candidate Set Size":"70000","Recall":"93.5%"},"uses_additional_data":true}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}