Welcome to Minlong Peng’s Homepage

About Me

I am now working at the Cognitive Computing Lab, Baidu Inc., as an algorithm expert. I obtained a Ph.D. degree in the School of Computer Science at Fudan University (2021/06), advised by Prof. Qi Zhang and Prof. Xuanjing Huang. Prior to joining in FudanNLP, I obtained my bachelor degree from School of Automation, Sun Yat-Sen University in 2016. It’s my honor to have worked with Prof. Weishi Zheng from SYSU, Dr. Chin-Yew Lin and Dr. Jing Liu from Microsoft Research Asia, and Mr. Changlong Sun and Mr. Kang Le from Alibaba DAMO Academy.

Phone: (+86) 13641651562; Personal Email: pengminlong@outlook.com; Github; Google Scholar

News

• A learning project for Parallel Computing was published
  I implemented a parallel learning system from scratch based on pytorch, which supports data parallel and tensor parallel. This is a learning project suitable for junior researchers to get familar with parallel computing or researchers who want to implement a parallel system from the scratch instead of using public frameworks.

• A paper was accepted by ACL Findings
  This work, titled NormNet: Normalize Noun Phrases for More Robust NLP, is the early step of our exploaration to break down facts and reasoning. It seems that the reviewers didn’t notice this point :) but it is exactly what I actually want to do!

• A paper on backdoor attack was rejected by ICML but received quite postive recognition from some peers
  This work, titled Learning to Change Label: Addressing Backdoor Overfitting of Poison-Label Backdoor Attacks, introduces an easy-to-implement method to improve the stealthiness of backdoor attacks. I do it because I like simple but effective method.

• I am hiring interns for studying pre-trained language models
  I am conducting research related to pre-trained language models, with a focus on cross-lingual transfer of these models, for example, transferring a language model mainly trained on English corpus to Chinese. Please feel free to contact me if you are interested in this research direction. Master and Ph.D. candiates of Computer Science are preferred.

Research Interest

• Cross-Lingual of Pretrained Language Models: Transferring English Language Model (e.g., LLaMA) to Chinese.
• Robust Natural Language Processing: Improving the robutness of NLP models, especially the neural network based ones, to attacks and delicate variations.

Education

• Fudan University, Shanghai, China                         2016/09 – 2021/06
  Ph.D. in Computer Science and Technology

• Sun Yat-sen University, Guangzhou, China           2012/09 – 2016/06
  B.S. in Automation

Working Experience

• Baidu Inc., Beijing, China                                           2021/09 – Now
  Algorithm Researcher

• Alibaba DAMO Academy, Hangzhou, China           2020/04 – 2020/10
  Research Intern

• Microsoft Research Asia, Beijing, China                 2015/07 – 2016/08
  Intern in Knowledge Computing Group

Selected Publications

• Minlong Peng, Zidi Xiong, Mingming Sun, Ping Li, Label-Smoothed Backdoor Attack, CORR 2022.

• Xin Wang, Minlong Peng, Mingming Sun, Ping Li, OIE@OIA: an Adaptable and Efficient Open Information Extraction Framework, In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (long paper, ACL 2022).

• Minlong Peng, Qi Zhang, Weighed Domain-Invariant Representation Learning for Cross-domain Sentiment Analysis, In Proceedings of the 28th International Conference of Computational Linguistics (long paper, COLING 2020).

• Minlong Peng, Ruotian Ma, Qi Zhang, Lujun Zhao, Mengxi Wei, Changlong Sun, Xuanjing Huang, Toward Recognizing More Entity Types in NER: An Efficient Implementation using Only Entity Lexicons, In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (long paper, EMNLP 2020(Findings)).

• Ruotian Ma^*, Minlong Peng^*, Qi Zhang, Zhongyu Wei, Xuanjing Huang, Simplify the Usage of Lexicons in Chinese NER, In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (equal contribution, long paper, ACL 2020)

• Minlong Peng^*, Xiaoyu Xing^*, Qi Zhang, Jinlan Fu, Xuanjing Huang, Distantly Supervised Named Entity Recognition using Positive-Unlabeled Learning, In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (equal contribution, long paper, ACL 2019).

• Minlong Peng, Qi Zhang, Yu-gang Jiang, Xuanjing Huang, Cross-Domain Sentiment Classification with Target Domain Specific Information, In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (long paper, ACL 2018).

• Minlong Peng, Qi Zhang, Xiaoyu Xing, Tao Gui, Jinlan Fu, Xuanjing Huang, Learning Task-specific Representation for Novel Words in Sequence Labeling, In the 28th International Joint Conference on Artificial Intelligence (IJCAI 2019).

• Minlong Peng, Qi Zhang, Xiaoyu Xing, Tao Gui, Xuanjing Huang, Yu-Gang Jiang, Keyu Ding, Zhigang Chen, Trainable Undersampling for Class-Imbalance Learning, In Proceedings of the 33th AAAI Conference on Artificial Intelligence (AAAI 2019).

• Minlong Peng, Qiyuan Bian, Qi Zhang, Tao Gui, Jinlan Fu, Lanjun Zeng, and Xuanjing Huang, Model the Long-Term Post History for Hashtag Recommendation, In CCF International Conference on Natural Language Processing and Chinese Computing (outstanding paper, NLPCC, 2019).

• Minlong Peng, Yaosong Lin, Lanjun Zeng, Tao Gui, and Qi Zhang, Modeling the Long-Term Post History for Personalized Hashtag Recommendation, In China National Conference on Chinese Computational Linguistics (CCL, 2019).

• Lanjun Zeng, Minlong Peng, Yaqi Liu, Liaosha Xu, Zhongyu Wei, Xuanjing Huang, Personal Hashtag Recommendation based on Few-Shot Learning, Journal of Chinese Information Processing 2021.

• Minlong Peng, Qi Zhang, Address Instance-level Label Prediction in Multiple Instance Learning, In Arxiv Preprint 2019.

• Tao Gui, Qi Zhang, Jingjing Gong, Minlong Peng, Keyu Ding, Xuanjing Huang, Transferring from Formal Newswire Domain with Hypernet for Twitter POS Tagging, In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (long paper, EMNLP 2018).

• Tao Gui, Qi Zhang, Haoran Huang, Minlong Peng, Xuanjing Huang, Part-of-speech tagging for twitter with adversarial neural networks, In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing (long paper, EMNLP 2017).

Professional Activities

• PC Reviewer: AAAI2022, AAAI2021, ACL2020, EMNLP2020, CIKM2020, ACL2019

• SPC Reviewer: IJCAI2021, IJCAI2023

Honors & Awards

• 2021 Global Top 100 Chinese Rising Stars in Artificial Intelligence
• 2021 Shanghai Excellent Doctoral Graduates
• 2020 Baidu PhD Fellowship Nomination Awards (about 20/5,000 applicants worldwide)
• 2020 Intel Fellowship
• 2019 Microsoft Scholarship Nomination
• 2019 National Scholarship of Fudan University
• 2018 Special Grade Scholarship for Ph.D. Students of Fudan University (Sponsored by Huawei)