{"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/unified-language-model-pre-training-for","title":"Unified Language Model Pre-training for Natural Language Understanding and Generation","arxiv_id":"1905.03197","date":"2019-05-08","proceeding":"NeurIPS 2019 12","authors":["Li Dong","Nan Yang","Wenhui Wang","Furu Wei","Xiaodong Liu","Yu Wang","Jianfeng Gao","Ming Zhou","Hsiao-Wuen Hon"],"abstract":"This paper presents a new Unified pre-trained Language Model (UniLM) that can be fine-tuned for both natural language understanding and generation tasks. The model is pre-trained using three types of language modeling tasks: unidirectional, bidirectional, and sequence-to-sequence prediction. The unified modeling is achieved by employing a shared Transformer network and utilizing specific self-attention masks to control what context the prediction conditions on. UniLM compares favorably with BERT on the GLUE benchmark, and the SQuAD 2.0 and CoQA question answering tasks. Moreover, UniLM achieves new state-of-the-art results on five natural language generation datasets, including improving the CNN/DailyMail abstractive summarization ROUGE-L to 40.51 (2.04 absolute improvement), the Gigaword abstractive summarization ROUGE-L to 35.75 (0.86 absolute improvement), the CoQA generative question answering F1 score to 82.5 (37.1 absolute improvement), the SQuAD question generation BLEU-4 to 22.12 (3.75 absolute improvement), and the DSTC7 document-grounded dialog response generation NIST-4 to 2.67 (human performance is 2.65). The code and pre-trained models are available at https://github.com/microsoft/unilm.","url_abs":"https://arxiv.org/abs/1905.03197v3","url_pdf":"https://arxiv.org/pdf/1905.03197v3.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":"unified-language-model-pre-training-for","repo_url":"https://github.com/microsoft/unilm","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/KnightZhang625/BERT_TF","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/LeonZh0u/Chatbot","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/YunwenTechnology/Unilm","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/facebookresearch/data2vec_vision","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/jiaruncao/BioCopyMechanism","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/robinsongh381/unilm_pytorch_korean","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/uabinf/nlp-fall-2019-project-anuradha_shinjitha","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"unified-language-model-pre-training-for","repo_url":"https://github.com/fuqiang-git-hub/unilmv1-Paddle","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"paddle","reach":{"status":"ok"}}],"tasks":[{"task_slug":"abstractive-text-summarization","task_name":"Abstractive Text Summarization"},{"task_slug":"document-summarization","task_name":"Document Summarization"},{"task_slug":"generative-question-answering","task_name":"Generative Question Answering"},{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"natural-language-understanding","task_name":"Natural Language Understanding"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"question-generation","task_name":"Question Generation"},{"task_slug":"question-generation","task_name":"Question-Generation"},{"task_slug":"response-generation","task_name":"Response Generation"},{"task_slug":"text-generation","task_name":"Text Generation"},{"task_slug":"text-summarization","task_name":"Text Summarization"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bert","method_name":"BERT"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-linear-decay","method_name":"Linear Warmup With Linear Decay"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"},{"method_slug":"weight-decay","method_name":"Weight Decay"},{"method_slug":"wordpiece","method_name":"WordPiece"}],"datasets_introduced":[{"slug":"liu-et-al-corpus","name":"Liu et al. Corpus","full_name":"Liu et al. Corpus"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/abstractive-text-summarization-on-cnn-daily","task":"Abstractive Text Summarization","dataset":"CNN / Daily Mail","model":"UniLM","rank_in_archive_order":25,"of":53,"metrics":{"ROUGE-1":"43.08","ROUGE-2":"20.43","ROUGE-L":"40.34"},"uses_additional_data":true},{"leaderboard":"/sota/document-summarization-on-cnn-daily-mail","task":"Document Summarization","dataset":"CNN / Daily Mail","model":"UniLM (Abstractive Summarization)","rank_in_archive_order":13,"of":26,"metrics":{"ROUGE-1":"43.08","ROUGE-2":"20.43","ROUGE-L":"40.34"},"uses_additional_data":true},{"leaderboard":"/sota/generative-question-answering-on-coqa","task":"Generative Question Answering","dataset":"CoQA","model":"UniLM","rank_in_archive_order":2,"of":3,"metrics":{"F1-Score":"82.5"},"uses_additional_data":true},{"leaderboard":"/sota/question-generation-on-squad11","task":"Question Generation","dataset":"SQuAD1.1","model":"UniLM","rank_in_archive_order":8,"of":13,"metrics":{"BLEU-4":"22.78","METEOR":"25.1","ROUGE-L":"51.1"},"uses_additional_data":true},{"leaderboard":"/sota/text-summarization-on-gigaword","task":"Text Summarization","dataset":"GigaWord","model":"UniLM","rank_in_archive_order":18,"of":41,"metrics":{"ROUGE-1":"38.90","ROUGE-2":"20.05","ROUGE-L":"36.00"},"uses_additional_data":true}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1905.03197","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}