Yichen Wu / 吴一尘 (Pronounced: Ee-chen Woo)

City University of Hong Kong, Tat Chee Avenue, Kowloon Tong, Kowloon, HongKong SAR

Visiting at Harvard University, Allston, MA, US

Email: wuyichen.am97@gmail.com
[Google Scholar] [Github] [Linkedin]

Biography

I am currently a final-year Ph.D. student in the Department of Computer Science at City University of Hong Kong under the supervision of Prof. Ying Wei and Prof. Kede Ma. Before joining CityU, I received my M.Sc degree under the supervision of Prof. Deyu Meng from the School of Mathematics and Statistics, Xi'an Jiaotong University. Additionally, I hold dual B.S. degrees from Beijing Jiaotong University, where I majored in Mathematics and minored in Economics and Management.

I am now a visiting scholar in the Visual Computing Group, Harvard, supervised by Prof. Hanspeter Pfister (IEEE Fellow, ACM Fellow) and working closely with Dr. Wanhua Li. I look forward to connecting with everyone in Boston!

I am in the 2025 fall job market and actively seeking postdoctoral and industry opportunities. Feel free to reach out to me!

Research Interest

My research interests include continual learning, transfer learning, and meta-learning. Currently, I am focusing on designing algorithms to enhance the efficiency of LLMs in handling continual learning and exploring methods to utilize transformers within LLMs efficiently. Additionally, I am dedicated to investigating the following research topics:

Research Experiences

Harvard VCG Lab, MA, U.S.

Visiting Scholar, Sep. 2024 ~ Now

Supervised by Prof. Hanspeter Pfister (IEEE Fellow, ACM Fellow) and working closely with Dr. Wanhua Li.

Youtu Lab (Original Jarvis Research Center)@Tencent, Shenzhen, China

Research Intern, Sep. 2023 ~ April. 2024

Supervised by Prof. Yefeng Zheng (IEEE Fellow, AIMBE Fellow) and Dr. Hong Wang

AI Lab (Machine Learning Center)@Tencent, Shenzhen, China

Research Intern (Tencent Rhino Bird Elite Talent Program), Feb. 2021 ~ Aug. 2023

Supervised by Dr. Long-Kai Huang and Dr. Peilin Zhao

News

  • [09/2024] One paper has been accepted by IJCV.
  • [09/2024] Our Dual-quant about the LLM Quantization is accepted for an oral presentation at NeurIPS 2024!
  • [09/2024] Thrilled to have spent a year as a visiting scholar at Harvard!
  • [06/2024] Happy to receive the ICML Travel Award (2024)!
  • [05/2024] Honored to receive the ICLR 2024 Outstanding Honorable Mention Award !
  • [05/2024] Two papers related to continual learning are accepted by ICML 2024.
  • [04/2024] Happy to be selected as a DAAD AInet Fellow (2024)!
  • [01/2024] Our VR-MCL for Meta-Continual Learning is accepted in ICLR 2024 (oral, 1.2% of submissions)!
  • [01/2024] I have been invited to serve as a reviewer for CVPR2024.
  • [09/2023] I have been invited to serve as a reviewer for ICLR2024.
  • [06/2023] I have been invited to be a reviewer for TNNLS.
  • [07/2023] Our CBA for alleviating the recency bias in continual learning is accepted in ICCV 2023.
  • [01/2023] Our L2AC for imbalanced semi-supervised learning is accepted in ICLR 2023.
  • [10/2022] I am honored with the NeurIPS 2022 travel award!
  • [09/2022] Our ATU for task augmentation in meta-learning is accepted in NeurIPS 2022 (spotlight).
  • [09/2021] Begin my journey at the City University of Hong Kong under the supervision of Prof. Ying Wei.
  • [06/2021] I graduate from XJTU, with heartfelt thanks to my supervisor, Deyu Meng.
  • [05/2021] One paper is accepted in MICCAI 2021.
  • [03/2021] Our survey paper on single-image and video deraining is accepted in Science China Information Sciences.
  • [12/2020] Our MSLC for learning with noisy labels is accepted in AAAI 2021.
  • [12/2020] One paper is accepted in Knowledge-Based System.
  • [09/2018] Begin my journey in IID Lab, Xi'an Jiaotong University under the supervision of Prof. Deyu Meng.

Selected Publications [Full List] (^Co-first author; *Corresponding author)

 

DuQuant: Distributing Outliers via Dual Transformation Makes Stronger Quantized LLMs.

Haokun Lin^, Haobo Xu^, Yichen Wu^, Jingzhi Cui, Yingtao Zhang, Linzhan Mou, Linqi Song, Zhenan Sun*, Ying Wei*
Neural Information Processing Systems, (NeurIPS), 2024. (Oral)
[Paper] [Code] [Website]

 

Mitigating Catastrophic Forgetting in Online Continual Learning by Modeling Previous Task Interrelations via Pareto Optimization

Yichen Wu^, Hong Wang^, Peilin Zhao, Yefeng Zheng, Ying Wei*, Long-Kai Huang*
International Conference on Machine Learning (ICML), 2024.
[Paper] [Code]

 

Federated Continual Learning via Prompt-based Dual Knowledge Transfer

Hongming Piao^, Yichen Wu^, Dapeng Wu, Ying Wei*
International Conference on Machine Learning (ICML), 2024.
[Paper] [Code]

 

Meta Continual Learning Revisited: Implicitly Enhancing Online Hessian Approximation via Variance Reduction

Yichen Wu, Long-kai Huang*, Renzhen Wang, Deyu Meng, Ying Wei*.
International Conference on Learning Representation (ICLR), 2024.

(Outstanding Paper Award Honorable Mention / Oral, 0.23% acceptance rate)

[Paper] [Code]

 

Adversarial Task Up-sampling for Meta-learning

Yichen Wu, Long-Kai Huang*, Ying Wei*
Neural Information Processing Systems, (NeurIPS), 2022. (Spotlight)
[Paper] [Code]

 

Learning to Purify Noisy Labels via Meta Soft Label Corrector

Yichen Wu, Jun Shu, Qi Xie, Qian Zhao, Deyu Meng*
Association for the Advancement of Artificial Intelligence (AAAI), 2021.
[Paper] [Code]

 

CBA: Improving Online Continual Learning via Continual Bias Adaptor

Quanziang Wang, Renzhen Wang*, Yichen Wu, Xixi Jia, Qian Zhao, Deyu Meng*
International Conference on Computer Vision(ICCV), 2023.
[Paper] [Code]

 

Imbalanced Semi-supervised Learning with Bias Adaptive Classifier

Renzhen Wang, Xixi Jia, Quanziang Wang, Yichen Wu, Deyu Meng*
International Conference on Learning Representation(ICLR), 2023.
[Paper] [Code]

 

Survey on Rain Removal From Videos or A Single Image

Hong Wang, Yichen Wu, Minghan Li, Qian Zhao, Deyu Meng*
SCIENCE CHINA Information Sciences(SCIS), 2022.
[Paper] [Code]

 

Neighbor Matching for Semi-supervised Learning

Renzhen Wang, Yichen Wu, Huai Chen, Lisheng Wang, Deyu Meng*
Medical Image Computing and Computer Assisted Intervention (MICCAI), 2021.
[Paper] [Code]

Selected Honors

Talks

Services

Journal & Conference Reviewer:

Close Friends and Collaborators