Xiaosen Wang | 王晓森

alt text 

Xiaosen Wang is a researcher at Huawei currently. He received his master degree (advised by Kun He) from Huazhong University of Science and Technology, China in 2022. He recevied his Bachelor’s degree from Excellent Student ACM Class at School of Computer Science and Technology, HUST in 2019. He is generally interesetd in Machine Learning and Deep Learning, especially in Adversarial Examples in Computer Vision (CV) and Natural Language Processing (NLP).

I am always open to any cooperation. If you are interesting in working with me, just drop me an email!

Email | Github | Google Scholar
A compelete list of papers about adversarial examples | List of papers about transfer-based attacks | TransferAttack

News

  • 01/07/2024 - One paper was accepted by ECCV 2024.
  • 27/02/2024 - Two papers were accepted by CVPR 2024.
  • 12/17/2023 - We have released a transfer-based attack framework: TransferAttack.
  • 12/06/2023 - I was invited as a reviewer for ICML 2024.
  • 10/29/2023 - I was invited as a reviewer for CVPR 2024.
  • 22/09/2023 - Two papers were accepted by NeurIPS 2023.
  • 19/09/2023 - We have released A compelete list of papers about adversarial examples and will update it everyday.
  • 27/08/2023 - I was invited as a reviewer for ICLR 2024.
  • 22/08/2023 - One paper was accepted by BMVC 2023.
  • 26/07/2023 - One paper was accepted by ACM MM 2023.
  • 14/07/2023 - One paper was accepted by ICCV 2023.
  • 13/07/2023 - I will give a talk about transfer-based attack on KCon 2023.
  • 25/04/2023 - I was invited as a reviewer for BMVC 2023.
  • 06/02/2023 - I was invited as a reviewer for NeurIPS 2023.
  • 28/02/2023 - One paper was accepted by CVPR 2023.
  • 06/02/2023 - I was invited as a reviewer for ICCV 2023.
  • 18/01/2023 - I was invited as a reviewer for UAI 2023.
  • 10/01/2023 - I was invited as a reviewer for ICME 2023.
  • 11/02/2022 - I was invited as a reviewer for ICML 2023.
  • 11/02/2022 - I was invited as a reviewer for CVPR 2023.
  • 10/06/2022 - One paper was accepted by EMNLP 2022 Findings.
  • 07/04/2022 - One paper was accepted by ECCV 2022.
  • 05/21/2022 - I was awarded as an outstanding reviewer for CVPR 2022.
  • 05/16/2022 - Two papers were accepted by UAI 2022.
  • 03/31/2022 - I was invited as a reviewer for NeurIPS 2022.
  • 03/02/2022 - I was invited as a reviewer for ECCV 2022.
  • 12/30/2021 - I was invited as a reviewer for ICML 2022.
  • 11/09/2021 - I was invited as a reviewer for CVPR 2022.
  • 10/16/2021 - One paper was accepted by BMVC 2021.
  • 07/23/2021 - One paper was accepted by ICCV 2021.
  • 05/12/2021 - One paper was accepted by UAI 2021.
  • 04/19/2021 - I was invited as a reviewer for EMNLP 2021.
  • 04/15/2021 - I was admitted into Tencent Rhino Bird Elite Talent Training Program.
  • 03/01/2021 - One paper was accepted by CVPR 2021.
  • 12/25/2020 - I was invited to give a talk of my paper by biendata and paperweekly.
  • 12/02/2020 - One paper was accepted by AAAI 2021.

Publications

† corresponding author, * equal contribution

Bag of Tricks to Boost Adversarial Transferability Image Attack
Zeliang Zhang, Rongyi Zhu, Wei Yao, Xiaosen Wang†, Chenliang Xu
European Conference on Computer Vision (ECCV), 2024
[arXiv] [Code]

Boosting Adversarial Transferability by Block Shuffle and Rotation Image Attack
Kunyu Wang, Xuanran He, Wenxuan Wang, Xiaosen Wang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024
[arXiv] [Code]

MMA-Diffusion: MultiModal Attack on Diffusion Models Multi-Modal Attack
Yijun Yang, Ruiyuan Gao, Xiaosen Wang, Nan Xu, Qiang Xu
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024
[arXiv] [Code]

Rethinking the Backward Propagation for Adversarial Transferability Image Attack
Xiaosen Wang, Kangheng Tong, Kun He
Advances in Neural Information Processing Systems (NeurIPS), 2023
[arXiv] [Code] [Poster] [Slides]

Boosting Adversarial Transferability by Achieving Flat Local Maxima Image Attack
Zhijin Ge*, Hongying Liu*, Xiaosen Wang*, Fanhua Shang, Yuanyuan Liu
Advances in Neural Information Processing Systems (NeurIPS), 2023
[arXiv] [Code] [Poster] [Slides]

Diversifying the High-level Features for better Adversarial Transferability Oral Image Attack
Zhiyuan Wang, Zeliang Zhang, Siyuan Liang, Xiaosen Wang
British Machine Vision Conference (BMVC), 2023
[arXiv] [Code]

Improving the Transferability of Adversarial Examples with Arbitrary Style Transfer Image Attack
Zhijin Ge, Fanhua Shang, Hongying Liu, Yuanyuan Liu, Liang Wan, Wei Feng, Xiaosen Wang
ACM International Conference on Multimedia (ACM MM), 2023
[arXiv] [Code]

Structure Invariant Transformation for better Adversarial Transferability Image Attack
Xiaosen Wang, Zeliang Zhang, Jianping Zhang
IEEE/CVF International Conference on Computer Vision (ICCV), 2023
[arXiv] [Code] [Poster]

Improving the Transferability of Adversarial Samples by Path-Augmented Method Image Attack
Jianping Zhang, Jen-tse Huang, Wenxuan Wang, Yichen Li, Weibin Wu, Xiaosen Wang, Yuxin Su, Michael Lyu
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023
[arXiv] [Code]

TextHacker: Learning based Hybrid Local Search Algorithm for Text Hard-label Adversarial Attack Textual Attack
Zhen Yu*, Xiaosen Wang*, Wanxiang Che, Kun He
Findings of Conference on Empirical Methods in Natural Language Processing (EMNLP findings), 2022
[arXiv] [Code] [Poster] [Slides]

Triangle Attack: A Query-efficient Decision-based Adversarial Attack Image Attack
Xiaosen Wang, Zeliang Zhang, Kangheng Tong, Dihong Gong, Kun He, Zhifeng Li, Wei Liu
European Conference on Computer Vision (ECCV), 2022
[arXiv] [Code] [Poster] [Slides]

Detecting Textual Adversarial Examples through Randomized Substitution and Vote Textual Defense
Xiaosen Wang*, Yifeng Xiong*, Kun He
Conference on Uncertainty in Artificial Intelligence (UAI), 2022
[arXiv] [Code] [Poster]

Robust Textual Embedding against Word-level Adversarial Attacks Textual Defense
Yichen Yang*, Xiaosen Wang*, Kun He
Conference on Uncertainty in Artificial Intelligence (UAI), 2022
[arXiv] [Code] [Poster]

Boosting Adversarial Transferability through Enhanced Momentum Image Attack
Xiaosen Wang*, Jiadong Lin*, Han Hu, Jingdong Wang, Kun He
British Machine Vision Conference (BMVC), 2021
[arXiv] [Code]

Admix: Enhancing the Transferability of Adversarial Attacks Image Attack
Xiaosen Wang, Xuanran He, Jingdong Wang, Kun He
International Conference on Computer Vision (ICCV), 2021
[arXiv] [Code] [Poster] [Slides]

Natural Language Adversarial Defense through Synonym Encoding Textual Defense
Xiaosen Wang, Hao Jin, Yichen Yang, Kun He
Conference on Uncertainty in Artificial Intelligence (UAI), 2021
[arXiv] [Code] [Poster] [Slides]

Enhancing the Transferability of Adversarial Attacks through Variance Tuning Image Attack
Xiaosen Wang, Kun He
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021
[arXiv] [Code] [Poster] [Slides]

Adversarial Training with Fast Gradient Projection Method against Synonym Substitution based Text Attacks Textual Defense
Xiaosen Wang*, Yichen Yang*, Yihe Deng*, Kun He
AAAI Conference on Artificial Intelligence (AAAI), 2021
[arXiv] [Code] [Poster] [Slides]

A New Anchor Word Selection Method for the Separable Topic Discovery
Kun He, Wu Wang, Xiaosen Wang†, John E. Hopcroft
WIREs Data Mining and Knowledge Discovery
[arXiv]

Preprint

Generating Visually Realistic Adversarial Patch Image Attack
Xiaosen Wang, Kunyu Wang
arXiv preprint arXiv:2312.03030, 2023
[arXiv]

Rethinking Mixup for Improving the Adversarial Transferability Image Attack
Xiaosen Wang, Zeyuan Yin
arXiv preprint arXiv:2311.17087, 2023
[arXiv]

Improving Adversarial Transferability with Scheduled Step Size and Dual Example Image Attack
Zeliang Zhang, Peihan Liu, Xiaosen Wang†, Chenliang Xu†
arXiv preprint arXiv:62301.12968, 2023
[arXiv]

I-PGD-AT: Efficient Adversarial Training via Imitating Iterative PGD Attack Image Defense
Xiaosen Wang, Bhavya Kailkhura, Krishnaram Kenthapadi, Bo Li
OpenReview: TEt7PsVZux6, 2021
[pdf]

Multi-stage Optimization based Adversarial Training Image Defense
Xiaosen Wang, Chuanbiao Song, Kun He, Liwei Wang
arXiv preprint arXiv:2106.15357, 2021
[arXiv]

AT-GAN: A Generative Attack Model for Adversarial Transferring on Generative Adversarial Nets Image Attack
Xiaosen Wang, Kun He, Chuanbiao Song, Liwei Wang, John E. Hopcroft
arXiv preprint arXiv:1904.07793, 2019
[arXiv]

Talks

Recent Progresses in Transfer-based Attack for Image Recognition [link] [slides]
Kcon. Aug 20, 2023

Adversarial Training in Textual Adversarial Attack and Defense [link]
AI Drive Organized by Biendata and Paperweekly. Jan 12, 2021

Vulnerability Detection for Android Applications [link]
XDef Natiaon Network and Information Security Conference. Nov 24, 2017

Selected Rewards

  • Outstanding Reviewer for CVPR, CVPR, May, 2022.
  • China National Scholarship, HUST, Sep. 2021.
  • Merit Master Student, HUST, Sep. 2020&2021.
  • BIGO Scholarship, HUST, Mar. 2021.
  • Tencent Rhino-BirdElite Talent Training Program, Tencent, May. 2021.
  • Shenzhen Stock Exchange Scholarship, HUST, Sep. 2020. (Only one student in CS, HUST is awarded by Shenzhen Stock Exchange each year.)
  • The CCF Elite Collegiate Award, CCF, Oct. 2018. (Only 3 students in HUST are awarded by China Computer Federation each year.)
  • Merit Student, HUST, Sep. 2018.
  • China National Scholarship, HUST, Sep. 2018.
  • China National Endeavor Scholarship, HUST, Sep. 2016.

Academic Service

  • Conference Reviewer: ICML, NeurIPS, ICLR, CVPR, ICCV, ECCV, ACL, EMNLP, UAI, ICME, BMVC, WACV, etc.
  • Journal Reviewer: TIFS, TMM, TCSVT, TNNLS, Neurocomputing, Pattern Recognition, IMAGE, etc.