Yipei Wang

About Me

I am currently a fifth-year Ph.D. candidate in the Elmore Family School of Electrical and Computer Engineering at Purdue University, advised by Dr. Xiaoqian Wang. Before coming to Purdue, I acquired an M.A. in Mathematics at University of Wisconsin-Madison. Prior to that, I obtained my B.S. in Computational Mathematics at Nanjing University.

My research focuses on building trustworthy machine learning systems. I work on developing reliable ML models, with an emphasis on rigorous evaluations of Explainable AI methods and their real-world applications. Recently, I have also been investigating robustness issues arising from natural distribution shifts, such as spurious correlations, and their relationship to adversarial robustness. Beyond these areas, I maintain broad interests in tackling challenges across various domains of machine learning.

Education

• 2020 - 2025 (Expected), Ph.D. in Computer Engineering, Purdue University
• 2018 - 2020, M.A. in Mathematics, University of Wisconsin-Madison
• 2015 - 2019, B.S. in Computational Mathematics, Nanjing University

Publications

• [ICML 2024]

  Wang, Y. & Wang, X. (2024) Benchmarking deletion metrics with the principled explanations. In The 41st International Conference on Machine Learning.

  [PDF] [code]
• [AISTATS 2024]

  Wang, Y. & Wang, X. (2024). On the effect of key factors in spurious correlation: a theoretical perspective. In International Conference on Artificial Intelligence and Statistics (pp. 3745-3753). PMLR.

  [PDF] [code]
• [ICLR 2024]

  Lu, S., Wang, Y., & Wang, X. (2024). Debiasing attention mechanism in transformer without demographics. In The 12th International Conference on Learning Representations.

  [PDF] [code]
• [ISBI 2024]

  Wang, Y., He, B., Risacher, S., Saykin, A., Yan, J., & Wang, X. (2024) Learning the irreversible progression trajectory of Alzheimer's disease. 21st IEEE international symposium on biomedical imaging.

  [PDF] [code]
• [NeurIPS 2022]

  Wang, Y., & Wang, X. (2022). "Why Not Other Classes?": Towards Class-Contrastive Back-Propagation Explanations. Advances in Neural Information Processing Systems, 35, 9085-9097.

  [PDF] [code]
• [NeurIPS 2021]

  Wang, Y., & Wang, X. (2021). Self-interpretable model with transformation equivariant interpretation. Advances in Neural Information Processing Systems, 34, 2359-2372.

  [PDF] [code]
• [CACAIE]

  Wang, Y., Hou, S., & Wang, X. (2021). Reinforcement learning‐based bird‐view automated vehicle control to avoid crossing traffic. Computer‐Aided Civil and Infrastructure Engineering, 36(7), 890-901.

  [PDF]

Preprint

• [arXiv]

  Wang, Y. & Wang, X. (2022). A unified study of machine learning explanation evaluation metrics. arXiv preprint arXiv:2203.14265.

  [PDF]

Contact

Email: wang4865@purdue.edu

LinkedIn: My Profile