Yisen Wang

Yisen Wang 王奕森
Assistant Professor, Ph.D. Advisor
School of Artificial Intelligence
Peking University


Room 2201, Science Building #2
Peking University, Beijing 100871, China


Email: yisen.wang AT pku DOT edu.cn 
Homepage: https://yisenwang.github.io/
Official Website: http://www.cis.pku.edu.cn/info/1084/1244.htm 

Official Website: http://www.ai.pku.edu.cn/info/1284/1643.htm

[Google Scholar] [Github]

[Students] [Recruitment Instructions]


Biography

I am now a Tenure-track Assistant Professor (Ph.D. Advisor) at Peking University. I am also a faculty member of ZERO Lab at Peking University led by Prof. Zhouchen Lin. I got my PhD degree from Department of Computer Science and Technology, Tsinghua University. I have visited Georgia Tech, USA, hosted by Prof. Le Song and Prof. Hongyuan Zha, and The University of Melbourne, Australia, hosted by Prof. James Bailey.

My research interest is broadly the representation learning from various types of data (unlabeled or noisy or adversarial data, structured data like graph, etc.). Specifically, we focus on theoretical and algorithmic approaches for adversarial machine learning, self-supervised/weakly supervised learning and graph learning.


Openings

We are recruiting highly motivated post-docs via Peking University Boya Postdoctoral Fellowship (Salary 350K+)

We are always actively recruiting Ph.D. students and interns! For Prospective Students, please read this note first!


Selected Publications

Equal  Contribution;  Corresponding  Author; [Full List])

  1. How Mask Matters: Towards Theoretical Understandings of Masked Autoencoders [PDF] [Code]
    Qi Zhang*, Yifei Wang*, Yisen Wang#
    Neural Information Processing Systems (NeurIPS 2022), 2022 (Spotlight, Top 5%)

  2. When Adversarial Training Meets Vision Transformers: Recipes from Training to Architecture [PDF] [Code]
    Yichuan Mo, Dongxian Wu, Yifei Wang, Yiwen Guo, Yisen Wang#
    Neural Information Processing Systems (NeurIPS 2022), 2022 (Spotlight, Top 5%)

  3. Improving Out-of-Distribution Robustness by Adversarial Training with Structured Priors [PDF] [Code]
    Qixun Wang*, Yifei Wang*, Hong Zhu, Yisen Wang#
    Neural Information Processing Systems (NeurIPS 2022), 2022 (Spotlight, Top 5%)

  4. Certified Adversarial Robustness Under the Bounded Support Set [PDF]
    Yiwen Kou, Qinyuan Zheng, Yisen Wang#
    International Conference on Machine Learning (ICML 2022), 2022

  5. CerDEQ: Certifiable Deep Equilibrium Model [PDF]
    Mingjie Li, Yisen Wang, Zhouchen Lin
    International Conference on Machine Learning (ICML 2022), 2022

  6. Optimization-Induced Graph Implicit Nonlinear Diffusion [PDF] [Code]
    Qi Chen, Yifei Wang, Yisen Wang, Jiansheng Yang, Zhouchen Lin
    International Conference on Machine Learning (ICML 2022), 2022

  7. G2CN: Graph Gaussian Convolution Networks with Concentrated Graph Filters [PDF] [Code]
    Mingjie Li, Xiaojun Guo, Yifei Wang, Yisen Wang, Zhouchen Lin
    International Conference on Machine Learning (ICML 2022), 2022

  8. Self-ensemble Adversarial Training for Improved Robustness [PDF] [Code]
    Hongjun Wang, Yisen Wang#
    International Conference on Learning Representations (ICLR 2022), 2022

  9. A Unified Contrastive Energy-based Model for Understanding the Generative Ability of Adversarial Training [PDF] [Code]
    Yifei Wang, Yisen Wang#, Jiansheng Yang, Zhouchen Lin
    International Conference on Learning Representations (ICLR 2022), 2022

  10. Chaos is a Ladder: A New Theoretical Understanding of Contrastive Learning via Augmentation Overlap [PDF] [Code]
    Yifei Wang*, Qi Zhang*, Yisen Wang#, Jiansheng Yang, Zhouchen Lin
    International Conference on Learning Representations (ICLR 2022), 2022

  11. Optimization inspired Multi-Branch Equilibrium Models [PDF] [Code]
    Mingjie Li, Yisen Wang, Xingyu Xie, Zhouchen Lin
    International Conference on Learning Representations (ICLR 2022), 2022

  12. Training High-Performance Low-Latency Spiking Neural Networks by Differentiation on Spike Representation [PDF] [Code]
    Qingyan Meng, Mingqing Xiao, Shen Yan, Yisen Wang, Zhouchen Lin, Zhiquan Luo
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2022), 2022

  13. Clustering Effect of (Linearized) Adversarial Robust Models [PDF] [Code]
    Yang Bai*, Xin Yan*, Yong Jiang, Shu-Tao Xia#, Yisen Wang#
    Neural Information Processing Systems (NeurIPS 2021), 2021 (Spotlight, Top 3%)

  14. Residual Relaxation for Multi-view Representation Learning [PDF]
    Yifei Wang, Zhengyang Geng, Feng Jiang, Chuming Li, Yisen Wang#, Jiansheng Yang, Zhouchen Lin
    Neural Information Processing Systems (NeurIPS 2021), 2021

  15. Dissecting the Diffusion Process in Linear Graph Convolutional Networks [PDF] [Code]
    Yifei Wang, Yisen Wang#, Jiansheng Yang, Zhouchen Lin
    Neural Information Processing Systems (NeurIPS 2021), 2021

  16. Adversarial Neuron Pruning Purifies Backdoored Deep Models [PDF] [Code]
    Dongxian Wu, Yisen Wang#
    Neural Information Processing Systems (NeurIPS 2021), 2021

  17. Moire Attack (MA): A New Potential Risk of Screen Photos [PDF] [Code]
    Dantong Niu, Guo Ruohao, Yisen Wang#
    Neural Information Processing Systems (NeurIPS 2021), 2021

  18. A Unified Game-Theoretic Interpretation of Adversarial Robustness [PDF] [Code]
    Jie Ren*, Die Zhang*, Yisen Wang*, Lu Chen, Zhanpeng Zhou, Yiting Chen, Xu Cheng, Xin Wang, Meng Zhou, Jie Shi, Quanshi Zhang
    Neural Information Processing Systems (NeurIPS 2021), 2021

  19. Exploring Architectural Ingredients of Adversarially Robust Deep Neural Networks [PDF] [Code]
    Hanxun Huang, Yisen Wang, Sarah Monazam Erfani, Quanquan Gu, James Bailey, Xingjun Ma
    Neural Information Processing Systems (NeurIPS 2021), 2021

  20. Finding Optimal Tangent Points for Reducing Distortions of Hard-label Attacks [PDF] [Code]
    Chen Ma, Xiangyu Guo, Li Chen, Jun-Hai Yong, Yisen Wang
    Neural Information Processing Systems (NeurIPS 2021), 2021

  21. On Training Implicit Models [PDF]
    Zhengyang Geng, Xin-Yu Zhang, Shaojie Bai, Yisen Wang, Zhouchen Lin
    Neural Information Processing Systems (NeurIPS 2021), 2021

  22. Gauge Equivariant Transformer [PDF]
    Lingshen He, Yiming Dong, Yisen Wang, Dacheng Tao, Zhouchen Lin
    Neural Information Processing Systems (NeurIPS 2021), 2021

  23. Training Feedback Spiking Neural Networks by Implicit Differentiation on the Equilibrium State [PDF] [Code]
    Mingqing Xiao, Qingyan Meng, Zongpeng Zhang, Yisen Wang, Zhouchen Lin
    Neural Information Processing Systems (NeurIPS 2021), 2021 (Spotlight, Top 3%)

  24. Reparameterized Sampling for Generative Adversarial Networks [PDF] [Code] [Award]
    Yifei Wang, Yisen Wang#, Jiansheng Yang, Zhouchen Lin
    European Conference on Machine Learning (ECML 2021), 2021 (Best (Student) Machine Learning Paper Award)

  25. Demystifying Adversarial Training via A Unified Probabilistic Framework [PDF] [Award]
    Yifei Wang, Yisen Wang#, Jiansheng Yang, Zhouchen Lin
    International Conference on Machine Learning Workshop (ICML Workshop 2021), 2021 (Silver Best Paper Award)

  26. Can Subnetwork Structure be the Key to Out-of-Distribution Generalization? [PDF]
    Dinghuai Zhang, Kartik Ahuja, Yilun Xu, Yisen Wang, Aaron Courville
    International Conference on Machine Learning (ICML 2021), 2021 (Long Talk, Top 3%)

  27. Leveraged Weighted Loss for Partial Label Learning [PDF] [Code]
    Hongwei Wen*, Jingyi Cui*, Hanyuan Hang, Jiabin Liu#, Yisen Wang#, Zhouchen Lin#
    International Conference on Machine Learning (ICML 2021), 2021 (Long Talk, Top 3%)

  28. GBHT: Gradient Boosting Histogram Transform for Density Estimation [PDF]
    Jingyi Cui*, Hanyuan Hang*, Yisen Wang#, Zhouchen Lin
    International Conference on Machine Learning (ICML 2021), 2021 

  29. Analysis and Applications of Class-wise Robustness in Adversarial Training [PDF]
    Qi Tian, Kun Kuang#, Kelu Jiang, Fei Wu, Yisen Wang#
    ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2021), 2021 

  30. Unlearnable Examples: Making Personal Data Unexploitable [PDF] [Code] [MIT Technology Review]
    Hanxun Huang, Xingjun Ma#, Sarah Monazam Erfani, James Bailey, Yisen Wang#
    International Conference on Learning Representations (ICLR 2021), 2021 (Spotlight, Top 4%)

  31. Improving Adversarial Robustness via Channel-wise Activation Suppressing [PDF] [Code]
    Yang Bai*, Yuyuan Zeng*, Yong Jiang, Shu-Tao Xia#, Xingjun Ma, Yisen Wang#
    International Conference on Learning Representations (ICLR 2021), 2021 (Spotlight, Top 4%)

  32. A Unified Approach to Interpreting and Boosting Adversarial Transferability [PDF] [Code]
    Xin Wang*, Jie Ren*, Shuyun Lin, Xiangming Zhu, Yisen Wang, Quanshi Zhang
    International Conference on Learning Representations (ICLR 2021), 2021

  33. Adversarial Weight Perturbation Helps Robust Generalization [PDF] [Code] [NeurIPS MeetUp Video]
    Dongxian Wu, Shu-Tao Xia, Yisen Wang#
    Neural Information Processing Systems (NeurIPS 2020), 2020 

  34. Normalized Loss Functions for Deep Learning with Noisy Labels [PDF] [Code]
    Xingjun Ma*, Hanxun Huang*, Yisen Wang#, Simone Romano, Sarah Erfani and James Bailey
    International Conference on Machine Learning (ICML 2020), 2020 

  35. Improving Adversarial Robustness Requires Revisiting Misclassified Examples [PDF] [Code]
    Yisen Wang*, Difan Zou*, Jinfeng Yi, James Bailey, Xingjun Ma, Quanquan Gu
    International Conference on Learning Representations (ICLR 2020), Addis Ababa, Ethiopia, 2020 

  36. Skip Connections Matter: On the Transferability of Adversarial Examples Generated with ResNets [PDF] [Code]
    Dongxian Wu, Yisen Wang#, Shu-Tao Xia, James Bailey, Xingjun Ma
    International Conference on Learning Representations (ICLR 2020), Addis Ababa, Ethiopia, 2020 (Spotlight)

  37. Adversarial Camouflage: Hiding Physical-World Attacks with Natural Styles [PDF] [Code]
    Ranjie Duan, Xingjun Ma, Yisen Wang, James Bailey, Kai Qin, Yun Yang
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2020), Seattle, USA, 2020 

  38. Improving Query Efficiency of Black-box Adversarial Attack [PDF] [Code]
    Yang Bai*, Yuyuan Zeng*, Yong Jiang#, Yisen Wang#, Shu-Tao Xia, Weiwei Guo
    European Conference on Computer Vision (ECCV 2020), 2020 

  39. Understanding Adversarial Attacks on Deep Learning Based Medical Image Analysis Systems [PDF]
    Xingjun Ma, Yuhao Niu, Lin Gu, Yisen Wang, Yitian Zhao, James Bailey, Feng Lu
    Pattern Recognition (PR), 2020

  40. On the Convergence and Robustness of Adversarial Training [PDF] [Code]
    Yisen Wang*, Xingjun Ma*, James Bailey, Jinfeng Yi, Bowen Zhou, Quanquan Gu
    International Conference on Machine Learning (ICML 2019), Long Beach, USA, 2019 (Long Talk)

  41. Symmetric Cross Entropy for Robust Learning with Noisy Labels [PDF] [Code]
    Yisen Wang*, Xingjun Ma*, Zaiyi Chen, Yuan Luo, Jinfeng Yi, James Bailey
    International Conference on Computer Vision (ICCV 2019), Seoul, Korea, 2019 

  42. Hilbert-Based Generative Defense for Adversarial Examples [PDF]
    Yang Bai*, Yan Feng*, Yisen Wang#, Shu-Tao Xia, Yong Jiang#
    International Conference on Computer Vision (ICCV 2019), Seoul, Korea, 2019 

  43. Learning Deep Hidden Nonlinear Dynamics from Aggregate Data [PDF][Appendix]
    Yisen Wang, Bo Dai, Lingkai Kong, Sarah Monazam Erfani, James Bailey, Hongyuan Zha
    Conference on Uncertainty in Artificial Intelligence (UAI 2018), California, USA, 2018

  44. Dimensionality-Driven Learning with Noisy Labels [PDF] [Code]
    Xingjun Ma*, Yisen Wang*, Michael E. Houle, Shuo Zhou, Sarah Monazam Erfani, Shu-Tao Xia, Sudanthi Wijewickrema, James Bailey
    International Conference on Machine Learning (ICML 2018), Stockholm, Sweden, 2018 (Long Talk)

  45. Iterative Learning with Open-set Noisy Labels [PDF] [Code]
    Yisen Wang, Weiyang Liu, Xingjun Ma, James Bailey, Hongyuan Zha, Le Song, Shu-Tao Xia
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2018), Salt Lake City, USA, 2018 (Spotlight)

  46. Decoupled Networks [PDF] [Code]
    Weiyang Liu, Zhen Liu, Zhiding Yu, Bo Dai, Rongmei Lin, Yisen Wang, James Rehg, Le Song
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2018), Salt Lake City, USA, 2018 (Spotlight)

  47. Characterizing Adversarial Subspaces Using Local Intrinsic Dimensionality [PDF] [Code]
    Xingjun Ma, Bo Li, Yisen Wang, Sarah M. Erfani, Sudanthi Wijewickrema, Michael E. Houle, Grant Schoenebeck, Dawn Song, James Bailey
    International Conference on Learning Representations (ICLR 2018), Vancouver, BC, Canada, 2018 (Oral)

  48. A Novel Consistent Random Forest Framework: Bernoulli Random Forests [PDF]
    Yisen Wang, Shu-Tao Xia, Qingtao Tang, Jia Wu, Xingquan Zhu
    IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2017


Academic Service

Talks


Awards