Jingwen Ye

Welcome to my homepage! I'm Jingwen Ye, currently a research fellowship at Learning and Vision Lab, National University of Singapore, working with Prof. Xinchao Wang. I received my PhD from Zhejiang University, studying computer vision under Prof. Mingli Song and Chun Chen. Prior to that, I got my B. Eng. Degree from Dalian University of Technology.

My current research interests are mainly about Privacy-aware Transfer Learning and Effective Model Reusing. Specially, I focus on the privacy issues on the AIGC models. Also I investigate deeper with knowledge distillation and amalgamation techniques to improve the performance of the multi-task networks.

Email  /  CV  /  Google Scholar  /  Github  /  Twitter  /  Linkedin

The world is full of magic things, patiently waiting for our senses to grow sharper. -- W.B. Yeats
profile photo


News & Invited Talk

- Apr. 2024. I was invited to give a guest lecture on “Efficiency- and Privacy- Related Model Reuse” at NUS.

- Mar. 2024. My two first-author papers were accepted by CVPR 2024

- Dec. 2022. My paper was voted Best Paper Award, IEEE VCIP 2022

- Aug. 2022. I was invited to give a talk on Privacy-related Transfer Learning to Shanghai AI lab.

- Nov 2022. Our Paper received a nomination for the Paper Award at NeurIPS 2022.

- June 2021. I graduated with a Ph.D. from Zhejiang University and was honored as an Graduate of Zhejiang Province.

- Feb 2020. I was invited to give a presentation on How To Plan Your Research Life During the COVID-19.

- Oct. 2019. I accepted an interview by the official media of the school and had My Personal Column published.


Selected Publications
Distilled Datamodel with Reverse Gradient Matching

Jingwen Ye, Ruonan Yu, Songhua Liu, Xinchao Wang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024
[paper]

We introduce an efficient framework for assessing data impact, comprising offline training and online evaluation stages.

Ungeneralizable Examples

Jingwen Ye, Xinchao Wang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024
[paper]

UGEs exhibit learnability for authorized users while maintaining unlearnability for potential hackers.

Mutual-modality Adversarial Attack with Semantic Perturbation

Jingwen Ye, Ruonan Yu, Songhua Liu, Xinchao Wang
AAAI Conference on Artificial Intelligence (AAAI), 2024
[paper]

We propose a novel approach that generates adversarial attacks in a mutual-modality optimization scheme.

Partial Network Cloning

Jingwen Ye, Songhua Liu, Xinchao Wang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023
[paper] [code] [video] [blog]

PNC conducts partial parametric "cloning" from a source network and then injects the cloned module to the target, without modifying its parameters.

Learning with Recoverable Forgetting

Jingwen Ye, Yifan Fu, Jie Song, Xingyi Yang, Songhua Liu, Xin Jin, Mingli Song, Xinchao Wang
European Conference on Computer Vision(ECCV), 2022
[paper] [code] [blog]

LIRF explicitly handles the task- or sample-specific knowledge removal and recovery.

Safe Distillation Box

Jingwen Ye, Yining Mao, Jie Song, Cheng Jing Xinchao Wang
d (AAAI), 2022
[paper] [code] [video] [blog]

SDB allows us to wrap a pre-trained model in a virtual box for intellectual property protection.



Work Experience
Research Fellow,  National University of Singapore

- Work in LVLab, Department of Electrical and Computer Engineering.
- Advise Prof. Xinchao Wang on privacy-related machine learning, effective model reuse, and dataset condensation.
Oct. 2021 -- Present
Research Intern,  Alibaba Group

- Developed a human matting method that was successfully applied in the Taobao APP, resulting in a first-round engagement of 339,187 PV and 82,210 UV.
Oct. 2017 -- Mar. 2019
Research Intern,&Alibaba-Zhejiang University Joint Institute

- Proposed the learning algorithm that supported the recommendation system.
- Received the honor of Outstanding Intern in 2019.

Jul. 2017 -- Sep. 2021
Education
Ph.D Student,  Zhejiang University,    (Sep. 2016 -- Jun. 2021)

College of Computer Science and Technology

Outstanding Graduate,   Advisor: Prof. Chun Chen and Prof. Mingli Song


B.Eng.,  Dalian University of Technology,   (Sep. 2012 -- Jun. 2016)

School of Information and Communication Engineering
Outstanding Graduate, Ranking: 1/35
Awards

Best Paper Award of International Conference on Visual Communications and Image Processing (2022)

Outstanding Graduate of Zhejiang Province (2021)

National Scholarship (top 2\%); Graduate of Merit/Triple A Graduate (2019 & 2020)

Excellent Intern of Alibaba-Zhejiang University Joint Research Institute of Frontier Technologise (2020)

Candidate of Zhu Kezhen Scholarship (top 1\%) (2019)

Most Valuable Academic Award of Doctoral Forum (2019)

Excellent Social Practice Individual Award (2018)

Award of Honor for Graduate (2017 & 2018)

Outstanding Graduate of Liaoning Province (2016)


Academic Service (Reviewer)

Journal Reviewer: TPAMI, TIP, SPM, TCYB, TCSVT, PR, TMLR, ...
Conference Reviewer: CVPR, ICCV, ECCV, ICLR, NeurIPS, ICML, AAAI, IJCAI, ...