{"type":"ScholarlyArticle","title":"Robust Fine-tuning via Perturbation and Interpolation from In-batch Instances","authors":[{"name":"Shoujie Tong","slug":"shoujie-tong","given":"Shoujie","family":"Tong","affiliation":"Key Laboratory of Computational Linguistics , Peking University"},{"name":"Qingxiu Dong","slug":"qingxiu-dong","given":"Qingxiu","family":"Dong","affiliation":"Key Laboratory of Computational Linguistics , Peking University"},{"name":"Damai Dai","slug":"damai-dai","given":"Damai","family":"Dai","affiliation":"Key Laboratory of Computational Linguistics , Peking University; Tencent Cloud Xiaowei"},{"name":"Yifan Song","slug":"yifan-song","given":"Yifan","family":"Song","affiliation":"Key Laboratory of Computational Linguistics , Peking University"},{"name":"Tianyu Liu","slug":"tianyu-liu","given":"Tianyu","family":"Liu","affiliation":"Key Laboratory of Computational Linguistics , Peking University"},{"name":"Baobao Chang","slug":"baobao-chang","given":"Baobao","family":"Chang","affiliation":"Key Laboratory of Computational Linguistics , Peking University"},{"name":"Zhifang Sui","slug":"zhifang-sui","given":"Zhifang","family":"Sui","affiliation":"Key Laboratory of Computational Linguistics , Peking University"}],"abstract":"Fine-tuning pretrained language models (PLMs) on downstream tasks has become common practice in natural language processing. However, most of the PLMs are vulnerable, e.g., they are brittle under adversarial attacks or imbalanced data, which hinders the application of the PLMs on some downstream tasks, especially in safe-critical scenarios. In this paper, we propose a simple yet effective fine-tuning method called Match-Tuning to force the PLMs to be more robust. For each instance in a batch, we involve other instances in the same batch to interact with it. To be specific, regarding the instances with other labels as a perturbation, Match-Tuning makes the model more robust to noise at the beginning of training. While nearing the end, Match-Tuning focuses more on performing an interpolation among the instances with the same label for better generalization. Extensive experiments on various tasks in GLUE benchmark show that Match-Tuning consistently outperforms the vanilla fine-tuning by 1.64 scores. Moreover, Match-Tuning exhibits remarkable robustness to adversarial attacks and data imbalance.","doi":"10.24963/ijcai.2022/610","url":"https://ijcai.org/proceedings/2022/610","pdf":"https://ijcai.org/proceedings/2022/0610.pdf","markdown":"https://ijcai.org/proceedings/2022/610.md","bibtex":"https://ijcai.org/proceedings/2022/bibtex/610","volume":{"slug":"2022","title":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2022","url":"https://ijcai.org/proceedings/2022/","year":2022,"location":"Vienna","dates":"23-29 July 2022","editors":"Luc De Raedt","isbn":"978-1-956792-00-3","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2022"},"section":"Natural Language Processing","track":"Main Track","number":610,"pages":{"first":4397,"last":4403},"keywords":["Natural Language Processing: Text Classification","Machine Learning: Classification"],"videos":[{"label":"Video #1","url":"https://slideslive.com/38986875","length_sec":87,"provider":"slideslive"},{"label":"Video #2","url":"https://slideslive.com/38985018","length_sec":316,"provider":"slideslive"}],"references":[{"position":1,"raw":"[Aghajanyan et al., 2021] Armen Aghajanyan, Akshat Shri- vastava, Anchit Gupta, Naman Goyal, Luke Zettlemoyer, and Sonal Gupta. Better fine-tuning by reducing represen- tational collapse. In International Conference on Learning Representations (ICLR), 2021.","title":"Better fine-tuning by reducing representational collapse","authors":"Armen Aghajanyan, Akshat Shrivastava, Anchit Gupta, Naman Goyal, Luke Zettlemoyer, Sonal Gupta","year":"2021","venue":"International Conference on Learning Representations (ICLR)","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Branco et al., 2021] Ruben Branco, António Branco, João Rodrigues, and João Silva. Shortcutted commonsense: Data spuriousness in deep learning of commonsense rea- soning. In Proceedings of the 2021 Conference on Em- pirical Methods in Natural Language Processing, pages 1504-1521, 2021.","title":"Shortcutted commonsense: Data spuriousness in deep learning of commonsense reasoning","authors":"Ruben Branco, António Branco, João Rodrigues, João Silva","year":"2021","venue":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Chen et al., 2020] Sanyuan Chen, Yutai Hou, Yiming Cui, Wanxiang Che, Ting Liu, and Xiangzhan Yu. Recall and learn: Fine-tuning deep pretrained language models with less forgetting. In Bonnie Webber, Trevor Cohn, Yulan He, and Yang Liu, editors, EMNLP, 2020.","title":"Recall and learn: Fine-tuning deep pretrained language models with less forgetting","authors":"Sanyuan Chen, Yutai Hou, Yiming Cui, Wanxiang Che, Ting Liu, Xiangzhan Yu","year":"2020","venue":"EMNLP","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Clark et al., 2019] Christopher Clark, Mark Yatskar, and Luke Zettlemoyer. Don't Take the Easy Way Out: Ensem- ble Based Methods for Avoiding Known Dataset Biases. arXiv, 2019.","title":"Don't Take the Easy Way Out: Ensemble Based Methods for Avoiding Known Dataset Biases","authors":"Christopher Clark, Mark Yatskar, Luke Zettlemoyer","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Daumé III, 2007] Hal Daumé III. Frustratingly easy domain adaptation. In Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics (ACL), June 2007.","title":"Hal Daumé III. Frustratingly easy domain adaptation","authors":"Iii Daumé","year":"2007","venue":"Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics (ACL)","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Devlin et al., 2019] Jacob Devlin, Ming-Wei Chang, Ken- ton Lee, and Kristina Toutanova. BERT: Pre-training of deep bidirectional transformers for language understand- ing. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL- HLT), June 2019.","title":"BERT: Pre-training of deep bidirectional transformers for language understanding","authors":"Jacob Devlin, Ming-Wei Chang, Kenton Lee, Kristina Toutanova","year":"2019","venue":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT)","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Dinh et al., 2017] Laurent Dinh, Razvan Pascanu, Samy Bengio, and Yoshua Bengio. Sharp minima can general- ize for deep nets. In International Conference on Machine Learning, pages 1019-1028. PMLR, 2017.","title":"Sharp minima can generalize for deep nets","authors":"Laurent Dinh, Razvan Pascanu, Samy Bengio, Yoshua Bengio","year":"2017","venue":"International Conference on Machine Learning","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Fan et al., 2021] Lijie Fan, Sijia Liu, Pin-Yu Chen, Gaoyuan Zhang, and Chuang Gan. When does contrastive learning preserve adversarial robustness from pretraining to finetuning? Advances in Neural Information Process- ing Systems, 34, 2021.","title":"When does contrastive learning preserve adversarial robustness from pretraining to finetuning?","authors":"Lijie Fan, Sijia Liu, Pin-Yu Chen, Gaoyuan Zhang, Chuang Gan","year":"2021","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Gao et al., 2021] Tianyu Gao, Xingcheng Yao, and Danqi Chen. Simcse: Simple contrastive learning of sentence embeddings. arXiv preprint arXiv:2104.08821, 2021.","title":"Simcse: Simple contrastive learning of sentence embeddings","authors":"Tianyu Gao, Xingcheng Yao, Danqi Chen","year":"2021","venue":null,"doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Hochreiter and Schmidhuber, 1995] Sepp Hochreiter and Jürgen Schmidhuber. Simplifying neural nets by discov- ering flat minima. In Advances in neural information pro- cessing systems, pages 529-536, 1995.","title":"Simplifying neural nets by discovering flat minima","authors":"Schmidhuber Hochreiter, Sepp Hochreiter, Jürgen Schmidhuber","year":"1995","venue":"Advances in neural information processing systems","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Houlsby et al., 2019] Neil Houlsby, Andrei Giurgiu, Stanis- law Jastrzebski, Bruna Morrone, Quentin De Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for NLP. In ICML 2019, 2019.","title":"Parameter-efficient transfer learning for NLP","authors":"Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin De Laroussilhe, Andrea Gesmundo, Mona Attariyan, Sylvain Gelly","year":"2019","venue":"ICML 2019","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Hovy and Søgaard, 2015] Dirk Hovy and Anders Søgaard. Tagging performance correlates with author age. In ACL, pages 483-488, 2015.","title":"Tagging performance correlates with author age","authors":"Søgaard ; Hovy, Dirk Hovy, Anders Søgaard","year":"2015","venue":"ACL","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Jiang et al., 2020] Haoming Jiang, Pengcheng He, Weizhu Chen, Xiaodong Liu, Jianfeng Gao, and Tuo Zhao. SMART: Robust and efficient fine-tuning for pre-trained natural language models through principled regularized optimization. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), July 2020.","title":"SMART: Robust and efficient fine-tuning for pre-trained natural language models through principled regularized optimization","authors":"Haoming Jiang, Pengcheng He, Weizhu Chen, Xiaodong Liu, Jianfeng Gao, Tuo Zhao","year":"2020","venue":"Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL)","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Lee et al., 2020] Cheolhyoung Lee, Kyunghyun Cho, and Wanmo Kang. Mixout: Effective regularization to finetune large-scale pretrained language models. In ICLR, 2020.","title":"Mixout: Effective regularization to finetune large-scale pretrained language models","authors":"Lee","year":"2020","venue":"ICLR","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Liu et al., 2019a] Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. Roberta: A robustly optimized BERT pretraining approach. arXiv preprint, arXiv:1907.11692, 2019.","title":"Roberta: A robustly optimized BERT pretraining approach","authors":"Liu","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Liu et al., 2019b] Yufan Liu, Jiajiong Cao, Bing Li, Chun- feng Yuan, Weiming Hu, Yangxi Li, and Yunqiang Duan. Knowledge distillation via instance relationship graph. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7096-7104, 2019.","title":"Knowledge distillation via instance relationship graph","authors":"Liu","year":"2019","venue":"Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Sagawa et al., 2019] Shiori Sagawa, Pang Wei Koh, Tat- sunori B Hashimoto, and Percy Liang. Distributionally robust neural networks for group shifts: On the importance of regularization for worst-case generalization. arXiv preprint arXiv:1911.08731, 2019.","title":"Distributionally robust neural networks for group shifts: On the importance of regularization for worst-case generalization","authors":"Shiori Sagawa, Pang Wei Koh, Tatsunori B Hashimoto, Percy Liang","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Smith et al., 2018] Samuel L Smith, Pieter-Jan Kinder- mans, Chris Ying, and Quoc V Le. Don't decay the learn- ing rate, increase the batch size. In International Confer- ence on Learning Representations, 2018.","title":"Don't decay the learning rate, increase the batch size","authors":"L Samuel, Pieter-Jan Smith, Chris Kindermans, Quoc V Ying, Le","year":"2018","venue":"International Conference on Learning Representations","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Wang et al., 2019] Alex Wang, Amanpreet Singh, Julian Michael, Felix Hill, Omer Levy, and Samuel R. Bowman. GLUE: A multi-task benchmark and analysis platform for natural language understanding. In ICLR, 2019.","title":"GLUE: A multi-task benchmark and analysis platform for natural language understanding","authors":"Alex Wang, Amanpreet Singh, Julian Michael, Felix Hill, Omer Levy, Samuel R Bowman","year":"2019","venue":"ICLR","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Wang et al., 2021] Boxin Wang, Chejian Xu, Shuohang Wang, Zhe Gan, Yu Cheng, Jianfeng Gao, Ahmed Has- san Awadallah, and Bo Li. Adversarial glue: A multi-task benchmark for robustness evaluation of language models. arXiv preprint arXiv:2111.02840, 2021.","title":"Adversarial glue: A multi-task benchmark for robustness evaluation of language models","authors":"Boxin Wang, Chejian Xu, Shuohang Wang, Zhe Gan, Yu Cheng, Jianfeng Gao, Ahmed Hassan Awadallah, Bo Li","year":"2021","venue":null,"doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Xu et al., 2021] Runxin Xu, Fuli Luo, Zhiyuan Zhang, Chuanqi Tan, Baobao Chang, Songfang Huang, and Fei Huang. Raise a child in large language model: Towards effective and generalizable fine-tuning. In EMNLP, pages 9514-9528, 2021.","title":"Raise a child in large language model: Towards effective and generalizable fine-tuning","authors":"Xu","year":"2021","venue":"EMNLP","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Yao et al., 2021] Yu Yao, Tongliang Liu, Mingming Gong, Bo Han, Gang Niu, and Kun Zhang. Instance-dependent label-noise learning under a structural causal model. Ad- vances in Neural Information Processing Systems, 34, 2021.","title":"Instance-dependent label-noise learning under a structural causal model","authors":"Yu Yao, Tongliang Liu, Mingming Gong, Bo Han, Gang Niu, Kun Zhang","year":"2021","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Zhu et al., 2019] Chen Zhu, Yu Cheng, Zhe Gan, Siqi Sun, Tom Goldstein, and Jingjing Liu. Freelb: Enhanced adver- sarial training for natural language understanding. arXiv preprint arXiv:1909.11764, 2019.","title":"Freelb: Enhanced adversarial training for natural language understanding","authors":"Chen Zhu, Yu Cheng, Zhe Gan, Siqi Sun, Tom Goldstein, Jingjing Liu","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}