{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,3]],"date-time":"2026-11-03T13:46:24Z","timestamp":1793713584941,"version":"4.6.1"},"reference-count":62,"publisher":"Association for Computing Machinery (ACM)","issue":"1","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Proc. VLDB Endow."],"published-print":{"date-parts":[[2020,9]]},"abstract":"<jats:p>We present Ditto, a novel entity matching system based on pre-trained Transformer-based language models. We fine-tune and cast EM as a sequence-pair classification problem to leverage such models with a simple architecture. Our experiments show that a straight-forward application of language models such as BERT, DistilBERT, or RoBERTa pre-trained on large text corpora already significantly improves the matching quality and outperforms previous state-of-the-art (SOTA), by up to 29% of F1 score on benchmark datasets. We also developed three optimization techniques to further improve Ditto's matching capability. Ditto allows domain knowledge to be injected by highlighting important pieces of input information that may be of interest when making matching decisions. Ditto also summarizes strings that are too long so that only the essential information is retained and used for EM. Finally, Ditto adapts a SOTA technique on data augmentation for text to EM to augment the training data with (difficult) examples. This way, Ditto is forced to learn \"harder\" to improve the model's matching capability. The optimizations we developed further boost the performance of Ditto by up to 9.8%. Perhaps more surprisingly, we establish that Ditto can achieve the previous SOTA results with at most half the number of labeled data. Finally, we demonstrate Ditto's effectiveness on a real-world large-scale EM task. On matching two company datasets consisting of 789K and 412K records, Ditto achieves a high F1 score of 96.5%.<\/jats:p>","DOI":"10.14778\/3421424.3421431","type":"journal-article","created":{"date-parts":[[2020,10,28]],"date-time":"2020-10-28T01:15:11Z","timestamp":1603847711000},"page":"50-60","source":"Crossref","is-referenced-by-count":348,"title":["Deep entity matching with pre-trained language models"],"prefix":"10.14778","volume":"14","author":[{"given":"Yuliang","family":"Li","sequence":"first","affiliation":[{"name":"Megagon Labs"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jinfeng","family":"Li","sequence":"additional","affiliation":[{"name":"Megagon Labs"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yoshihiko","family":"Suhara","sequence":"additional","affiliation":[{"name":"Megagon Labs"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"AnHai","family":"Doan","sequence":"additional","affiliation":[{"name":"University of Wisconsin Madison"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wang-Chiew","family":"Tan","sequence":"additional","affiliation":[{"name":"Megagon Labs"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2020,10,27]]},"reference":[{"key":"e_1_2_1_1_1","doi-asserted-by":"publisher","DOI":"10.1109\/ICDE.2019.00114"},{"key":"e_1_2_1_2_1","unstructured":"Linkage Rohan Baxter Rohan Baxter Peter Christen etal 2003. A comparison of fast blocking methods for record. (2003). Linkage Rohan Baxter Rohan Baxter Peter Christen et al. 2003. A comparison of fast blocking methods for record. (2003)."},{"key":"e_1_2_1_3_1","volume-title":"SciBERT: A pretrained language model for scientific text. arXiv preprint arXiv:1903.10676","author":"Beltagy Iz","year":"2019","unstructured":"Iz Beltagy , Kyle Lo , and Arman Cohan . 2019. SciBERT: A pretrained language model for scientific text. arXiv preprint arXiv:1903.10676 ( 2019 ). Iz Beltagy, Kyle Lo, and Arman Cohan. 2019. SciBERT: A pretrained language model for scientific text. arXiv preprint arXiv:1903.10676 (2019)."},{"key":"e_1_2_1_4_1","doi-asserted-by":"publisher","DOI":"10.1145\/956750.956759"},{"key":"e_1_2_1_5_1","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00051"},{"key":"e_1_2_1_6_1","unstructured":"Ursin Brunner and Kurt Stockinger. 2020. Entity matching with transformer architectures-a step forward in data integration. In EDBT. Ursin Brunner and Kurt Stockinger. 2020. Entity matching with transformer architectures-a step forward in data integration. In EDBT."},{"key":"e_1_2_1_7_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P18-1224"},{"key":"e_1_2_1_8_1","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2011.127"},{"key":"e_1_2_1_9_1","volume-title":"Proc. BlackBoxNLP '19","author":"Clark Kevin","unstructured":"Kevin Clark , Urvashi Khandelwal , Omer Levy , and Christopher D. Manning . 2019. What Does BERT Look at? An Analysis of BERT's Attention . In Proc. BlackBoxNLP '19 . 276--286. Kevin Clark, Urvashi Khandelwal, Omer Levy, and Christopher D. Manning. 2019. What Does BERT Look at? An Analysis of BERT's Attention. In Proc. BlackBoxNLP '19. 276--286."},{"key":"e_1_2_1_10_1","doi-asserted-by":"publisher","DOI":"10.1145\/775047.775116"},{"key":"e_1_2_1_11_1","doi-asserted-by":"publisher","DOI":"10.1145\/2488388.2488415"},{"key":"e_1_2_1_12_1","unstructured":"Sanjib Das AnHai Doan Paul Suganthan G. C. Chaitanya Gokhale Pradap Konda Yash Govind and Derek Paulsen. [n.d.]. The Magellan Data Repository. https:\/\/sites.google.com\/site\/anhaidgroup\/projects\/data. Sanjib Das AnHai Doan Paul Suganthan G. C. Chaitanya Gokhale Pradap Konda Yash Govind and Derek Paulsen. [n.d.]. The Magellan Data Repository. https:\/\/sites.google.com\/site\/anhaidgroup\/projects\/data."},{"key":"e_1_2_1_13_1","volume-title":"Proc. NAACL-HLT '19","author":"Devlin Jacob","year":"2019","unstructured":"Jacob Devlin , Ming-Wei Chang , Kenton Lee , and Kristina Toutanova . 2019 . BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding . In Proc. NAACL-HLT '19 . 4171--4186. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proc. NAACL-HLT '19. 4171--4186."},{"key":"e_1_2_1_14_1","doi-asserted-by":"publisher","DOI":"10.5555\/3236187.3269461"},{"key":"e_1_2_1_15_1","doi-asserted-by":"publisher","DOI":"10.1145\/2588555.2594511"},{"key":"e_1_2_1_16_1","doi-asserted-by":"publisher","DOI":"10.1145\/2783258.2783396"},{"key":"e_1_2_1_17_1","doi-asserted-by":"publisher","DOI":"10.5555\/3367722.3367735"},{"key":"e_1_2_1_18_1","doi-asserted-by":"publisher","DOI":"10.1145\/2588555.2588576"},{"key":"e_1_2_1_19_1","doi-asserted-by":"publisher","DOI":"10.1145\/3357384.3360316"},{"key":"e_1_2_1_20_1","volume-title":"Thomas M\u00fcller, Francesco Piccinno, and Julian Martin Eisenschlos.","author":"Herzig Jonathan","year":"2020","unstructured":"Jonathan Herzig , Pawe\u0142 Krzysztof Nowak , Thomas M\u00fcller, Francesco Piccinno, and Julian Martin Eisenschlos. 2020 . TAPAS : Weakly Supervised Table Parsing via Pre-training . arXiv preprint arXiv:2004.02349 (2020). Jonathan Herzig, Pawe\u0142 Krzysztof Nowak, Thomas M\u00fcller, Francesco Piccinno, and Julian Martin Eisenschlos. 2020. TAPAS: Weakly Supervised Table Parsing via Pre-training. arXiv preprint arXiv:2004.02349 (2020)."},{"key":"e_1_2_1_21_1","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1997.9.8.1735"},{"key":"e_1_2_1_22_1","doi-asserted-by":"publisher","DOI":"10.14778\/2047485.2047487"},{"key":"e_1_2_1_23_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1586"},{"key":"e_1_2_1_24_1","doi-asserted-by":"publisher","DOI":"10.14778\/2994509.2994535"},{"key":"e_1_2_1_25_1","doi-asserted-by":"publisher","DOI":"10.14778\/1920841.1920904"},{"key":"e_1_2_1_26_1","doi-asserted-by":"crossref","first-page":"1234","DOI":"10.1093\/bioinformatics\/btz682","article-title":"BioBERT: a pre-trained biomedical language representation model for biomedical text mining","volume":"36","author":"Lee Jinhyuk","year":"2020","unstructured":"Jinhyuk Lee , Wonjin Yoon , Sungdong Kim , Donghyeon Kim , Sunkyu Kim , Chan Ho So , and Jaewoo Kang . 2020 . BioBERT: a pre-trained biomedical language representation model for biomedical text mining . Bioinformatics 36 , 4 (2020), 1234 -- 1240 . Jinhyuk Lee, Wonjin Yoon, Sungdong Kim, Donghyeon Kim, Sunkyu Kim, Chan Ho So, and Jaewoo Kang. 2020. BioBERT: a pre-trained biomedical language representation model for biomedical text mining. Bioinformatics 36, 4 (2020), 1234--1240.","journal-title":"Bioinformatics"},{"key":"e_1_2_1_27_1","doi-asserted-by":"publisher","DOI":"10.14778\/3421424.3421431"},{"key":"e_1_2_1_28_1","volume-title":"RoBERTa: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692","author":"Liu Yinhan","year":"2019","unstructured":"Yinhan Liu , Myle Ott , Naman Goyal , Jingfei Du , Mandar Joshi , Danqi Chen , Omer Levy , Mike Lewis , Luke Zettlemoyer , and Veselin Stoyanov . 2019. RoBERTa: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692 ( 2019 ). Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. RoBERTa: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692 (2019)."},{"key":"e_1_2_1_29_1","doi-asserted-by":"publisher","DOI":"10.1145\/3318464.3380597"},{"key":"e_1_2_1_30_1","doi-asserted-by":"publisher","DOI":"10.1145\/3366423.3380144"},{"key":"e_1_2_1_31_1","volume-title":"Proc. EMNLP '04","author":"Mihalcea Rada","year":"2004","unstructured":"Rada Mihalcea and Paul Tarau . 2004 . TextRank: Bringing order into text . In Proc. EMNLP '04 . 404--411. Rada Mihalcea and Paul Tarau. 2004. TextRank: Bringing order into text. In Proc. EMNLP '04. 404--411."},{"key":"e_1_2_1_32_1","doi-asserted-by":"publisher","DOI":"10.1145\/3183713.3196926"},{"key":"e_1_2_1_33_1","volume-title":"Blocking and Filtering Techniques for Entity Resolution: A Survey. arXiv preprint arXiv:1905.06167","author":"Papadakis George","year":"2019","unstructured":"George Papadakis , Dimitrios Skoutas , Emmanouil Thanos , and Themis Palpanas . 2019. Blocking and Filtering Techniques for Entity Resolution: A Survey. arXiv preprint arXiv:1905.06167 ( 2019 ). George Papadakis, Dimitrios Skoutas, Emmanouil Thanos, and Themis Palpanas. 2019. Blocking and Filtering Techniques for Entity Resolution: A Survey. arXiv preprint arXiv:1905.06167 (2019)."},{"key":"e_1_2_1_34_1","volume-title":"Proc. NeurIPS '19","author":"Paszke Adam","year":"2019","unstructured":"Adam Paszke , Sam Gross , Francisco Massa , Adam Lerer , James Bradbury , Gregory Chanan , Trevor Killeen , Zeming Lin , Natalia Gimelshein , Luca Antiga , 2019 . PyTorch: An imperative style, high-performance deep learning library . In Proc. NeurIPS '19 . 8024--8035. Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. 2019. PyTorch: An imperative style, high-performance deep learning library. In Proc. NeurIPS '19. 8024--8035."},{"key":"e_1_2_1_35_1","doi-asserted-by":"publisher","DOI":"10.5555\/1953048.2078195"},{"key":"e_1_2_1_36_1","doi-asserted-by":"publisher","DOI":"10.3115\/v1\/D14-1162"},{"key":"e_1_2_1_37_1","doi-asserted-by":"publisher","DOI":"10.1145\/3308560.3316609"},{"key":"e_1_2_1_38_1","doi-asserted-by":"publisher","DOI":"10.1145\/3132847.3132949"},{"key":"e_1_2_1_39_1","unstructured":"Alec Radford Jeff Wu Rewon Child David Luan Dario Amodei and Ilya Sutskever. 2019. Language Models are Unsupervised Multitask Learners. (2019). Alec Radford Jeff Wu Rewon Child David Luan Dario Amodei and Ilya Sutskever. 2019. Language Models are Unsupervised Multitask Learners. (2019)."},{"key":"e_1_2_1_40_1","first-page":"9","article-title":"Language models are unsupervised multitask learners","volume":"1","author":"Radford Alec","year":"2019","unstructured":"Alec Radford , Jeffrey Wu , Rewon Child , David Luan , Dario Amodei , and Ilya Sutskever . 2019 . Language models are unsupervised multitask learners . OpenAI Blog 1 , 8 (2019), 9 . Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. OpenAI Blog 1, 8 (2019), 9.","journal-title":"OpenAI Blog"},{"key":"e_1_2_1_41_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D19-1410"},{"key":"e_1_2_1_42_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D15-1044"},{"key":"e_1_2_1_43_1","volume-title":"Proc. EMC2 '19","author":"Sanh Victor","year":"2019","unstructured":"Victor Sanh , Lysandre Debut , Julien Chaumond , and Thomas Wolf . 2019 . DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter . In Proc. EMC2 '19 . Victor Sanh, Lysandre Debut, Julien Chaumond, and Thomas Wolf. 2019. DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter. In Proc. EMC2 '19."},{"key":"e_1_2_1_44_1","doi-asserted-by":"publisher","DOI":"10.1145\/775047.775087"},{"key":"e_1_2_1_45_1","doi-asserted-by":"publisher","DOI":"10.14778\/3149193.3149199"},{"key":"e_1_2_1_46_1","unstructured":"Spacy. [n.d.]. https:\/\/spacy.io\/api\/entityrecognizer. Spacy. [n.d.]. https:\/\/spacy.io\/api\/entityrecognizer."},{"key":"e_1_2_1_47_1","volume-title":"ERNIE: Enhanced representation through knowledge integration. arXiv preprint arXiv:1904.09223","author":"Sun Yu","year":"2019","unstructured":"Yu Sun , Shuohuan Wang , Yukun Li , Shikun Feng , Xuyi Chen , Han Zhang , Xin Tian , Danxiang Zhu , Hao Tian , and Hua Wu . 2019 . ERNIE: Enhanced representation through knowledge integration. arXiv preprint arXiv:1904.09223 (2019). Yu Sun, Shuohuan Wang, Yukun Li, Shikun Feng, Xuyi Chen, Han Zhang, Xin Tian, Danxiang Zhu, Hao Tian, and Hua Wu. 2019. ERNIE: Enhanced representation through knowledge integration. arXiv preprint arXiv:1904.09223 (2019)."},{"key":"e_1_2_1_48_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1452"},{"key":"e_1_2_1_49_1","doi-asserted-by":"publisher","DOI":"10.5555\/3295222.3295349"},{"key":"e_1_2_1_50_1","doi-asserted-by":"publisher","DOI":"10.14778\/2350229.2350263"},{"key":"e_1_2_1_51_1","doi-asserted-by":"publisher","DOI":"10.14778\/2021017.2021020"},{"key":"e_1_2_1_52_1","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2015.2468711"},{"key":"e_1_2_1_53_1","doi-asserted-by":"publisher","DOI":"10.1145\/3292500.3330989"},{"key":"e_1_2_1_54_1","unstructured":"WDC Product Data Corpus. [n.d.]. http:\/\/webdatacommons.org\/largescaleproductcorpus\/v2. WDC Product Data Corpus. [n.d.]. http:\/\/webdatacommons.org\/largescaleproductcorpus\/v2."},{"key":"e_1_2_1_55_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D19-1670"},{"key":"e_1_2_1_56_1","doi-asserted-by":"crossref","unstructured":"Thomas Wolf Lysandre Debut Victor Sanh Julien Chaumond Clement Delangue Anthony Moi Pierric Cistac Tim Rault R\u00e9mi Louf Morgan Funtowicz etal 2019. HuggingFace's Transformers: State-of-the-art Natural Language Processing. arXiv preprint arXiv:1910.03771 (2019). Thomas Wolf Lysandre Debut Victor Sanh Julien Chaumond Clement Delangue Anthony Moi Pierric Cistac Tim Rault R\u00e9mi Louf Morgan Funtowicz et al. 2019. HuggingFace's Transformers: State-of-the-art Natural Language Processing. arXiv preprint arXiv:1910.03771 (2019).","DOI":"10.18653\/v1\/2020.emnlp-demos.6"},{"key":"e_1_2_1_57_1","volume-title":"Unsupervised data augmentation. arXiv preprint arXiv:1904.12848","author":"Xie Qizhe","year":"2019","unstructured":"Qizhe Xie , Zihang Dai , Eduard Hovy , Minh-Thang Luong , and Quoc V Le. 2019. Unsupervised data augmentation. arXiv preprint arXiv:1904.12848 ( 2019 ). Qizhe Xie, Zihang Dai, Eduard Hovy, Minh-Thang Luong, and Quoc V Le. 2019. Unsupervised data augmentation. arXiv preprint arXiv:1904.12848 (2019)."},{"key":"e_1_2_1_58_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P17-1132"},{"key":"e_1_2_1_59_1","volume-title":"Proc. NeurIPS '19","author":"Yang Zhilin","year":"2019","unstructured":"Zhilin Yang , Zihang Dai , Yiming Yang , Jaime Carbonell , Russ R Salakhutdinov , and Quoc V Le . 2019 . XLNet: Generalized autoregressive pretraining for language understanding . In Proc. NeurIPS '19 . 5754--5764. Zhilin Yang, Zihang Dai, Yiming Yang, Jaime Carbonell, Russ R Salakhutdinov, and Quoc V Le. 2019. XLNet: Generalized autoregressive pretraining for language understanding. In Proc. NeurIPS '19. 5754--5764."},{"key":"e_1_2_1_60_1","volume-title":"TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data. arXiv preprint arXiv:2005.08314","author":"Yin Pengcheng","year":"2020","unstructured":"Pengcheng Yin , Graham Neubig , Wen-tau Yih, and Sebastian Riedel . 2020. TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data. arXiv preprint arXiv:2005.08314 ( 2020 ). Pengcheng Yin, Graham Neubig, Wen-tau Yih, and Sebastian Riedel. 2020. TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data. arXiv preprint arXiv:2005.08314 (2020)."},{"key":"e_1_2_1_61_1","volume-title":"Proc. ICLR '18","author":"Zhang Hongyi","year":"2018","unstructured":"Hongyi Zhang , Moustapha Cisse , Yann N Dauphin , and David Lopez-Paz . 2018 . mixup: Beyond empirical risk minimization . In Proc. ICLR '18 . Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez-Paz. 2018. mixup: Beyond empirical risk minimization. In Proc. ICLR '18."},{"key":"e_1_2_1_62_1","doi-asserted-by":"publisher","DOI":"10.1145\/3308558.3313578"}],"container-title":["Proceedings of the VLDB Endowment"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.14778\/3421424.3421431","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,10,10]],"date-time":"2023-10-10T18:57:18Z","timestamp":1696964238000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.14778\/3421424.3421431"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,9]]},"references-count":62,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2020,9]]}},"alternative-id":["10.14778\/3421424.3421431"],"URL":"https:\/\/doi.org\/10.14778\/3421424.3421431","relation":{},"ISSN":["2150-8097"],"issn-type":[{"value":"2150-8097","type":"print"}],"subject":[],"published":{"date-parts":[[2020,9]]}}}