{"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/ensembling-and-knowledge-distilling-of-large-1","title":"Ensembling and Knowledge Distilling of Large Sequence Taggers for Grammatical Error Correction","arxiv_id":"2203.13064","date":"2022-03-24","proceeding":"ACL 2022 5","authors":["Maksym Tarnavskyi","Artem Chernodub","Kostiantyn Omelianchuk"],"abstract":"In this paper, we investigate improvements to the GEC sequence tagging architecture with a focus on ensembling of recent cutting-edge Transformer-based encoders in Large configurations. We encourage ensembling models by majority votes on span-level edits because this approach is tolerant to the model architecture and vocabulary size. Our best ensemble achieves a new SOTA result with an $F_{0.5}$ score of 76.05 on BEA-2019 (test), even without pre-training on synthetic datasets. In addition, we perform knowledge distillation with a trained ensemble to generate new synthetic training datasets, \"Troy-Blogs\" and \"Troy-1BW\". Our best single sequence tagging model that is pretrained on the generated Troy-datasets in combination with the publicly available synthetic PIE dataset achieves a near-SOTA (To the best of our knowledge, our best single model gives way only to much heavier T5 model result with an $F_{0.5}$ score of 73.21 on BEA-2019 (test). The code, datasets, and trained models are publicly available).","url_abs":"https://arxiv.org/abs/2203.13064v1","url_pdf":"https://arxiv.org/pdf/2203.13064v1.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":"ensembling-and-knowledge-distilling-of-large-1","repo_url":"https://github.com/makstarnavskyi/gector-large","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"grammatical-error-correction","task_name":"Grammatical Error Correction"},{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"}],"methods":[{"method_slug":"adafactor","method_name":"Adafactor"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"glu","method_name":"Gated Linear Unit"},{"method_slug":"inverse-square-root-schedule","method_name":"Inverse Square Root Schedule"},{"method_slug":"knowledge-distillation","method_name":"Knowledge Distillation"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"sentencepiece","method_name":"SentencePiece"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"t5","method_name":"T5"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/grammatical-error-correction-on-bea-2019-test","task":"Grammatical Error Correction","dataset":"BEA-2019 (test)","model":"DeBERTa + RoBERTa + XLNet","rank_in_archive_order":7,"of":19,"metrics":{"F0.5":"76.05"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2203.13064","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}