PeopleofSuperintelligence

Researcher

Ke Xu

Also known as 许可

Ke Xu is a computer scientist at Shanghai Jiao Tong University whose research on digital media content security centers on video anomaly detection, gait recognition and adversarial attacks on video models. He studied at Shanghai Jiao Tong from 2009 and earned a doctorate there in 2019. His papers in IEEE Transactions on Multimedia in 2018, 2020 and 2021 covered video anomaly detection and skeleton-based gait recognition, and from 2022 he coauthored papers on adversarial attacks against video, face recognition and object tracking models. In 2022 he was runner-up in the Open Set Defense track of the AISafety challenge at CVPR 2022, and in 2024 he was among the completers of two projects that won first prizes, in the Shanghai Science and Technology Progress Award and the Chinese Society of Image and Graphics award. He is an associate researcher and doctoral supervisor in the School of Computer Science at Shanghai Jiao Tong, and his 2025 and 2026 coauthored papers cover jailbreaking text-to-image models, prompt injection against LLM agents and omni-modal temporal grounding.

Timeline

  1. 2009–2019

    Degrees at Shanghai Jiao Tong

    He studies at Shanghai Jiao Tong University from 2009 to 2013 for a bachelor's degree in engineering at the School of Information Security Engineering. He then studies from 2013 to 2019 for a doctorate in engineering at the School of Cyberspace Security, and graduates with his PhD in 2019.

  2. 2018–2020

    Video anomaly detection in TMM

    In 2018, he publishes Anomaly Detection Based on Stacked Sparse Coding with Intraframe Classification Strategy in IEEE Transactions on Multimedia, with Xinghao Jiang and Tanfeng Sun. In 2020, the journal publishes their Video Anomaly Detection and Localization Based on an Adaptive Intra-frame Classification Network.

  3. 2021

    Skeleton-based gait recognition

    In 2021, he publishes Gait Recognition based on Local Graphical Skeleton Descriptor with Pairwise Similarity Network in IEEE Transactions on Multimedia, with Xinghao Jiang and Tanfeng Sun. With Xingkai Zheng, Xirui Li, Jiang and Sun, he also coauthors a skeleton-based gait identification network built on a spatio-temporal graph convolutional network and tested on the CASIA-B dataset.

  4. 2022

    AISafety challenge runner-up

    In 2022, he is runner-up in the Open Set Defense track of the AISafety challenge at CVPR 2022. He also wins third prize in the CSIG FAT-AI 2022 Open Scene Face Combat Disguise Challenge.

  5. 2022–2025

    Adversarial attack papers

    In 2022, he coauthors Research on Video Adversarial Attack with Long Living Cycle at UAI, with Zeyu Zhao, Xinghao Jiang and Tanfeng Sun. His adversarial-attack papers continue with a face-recognition attack in IEEE Transactions on Information Forensics and Security in 2023 and an attack on visual object tracking in IEEE Transactions on Circuits and Systems for Video Technology in 2025.

  6. 2024

    Two first-prize science awards

    In 2024, he is among the completers of a project that wins a first prize in the Shanghai Science and Technology Progress Award, ranked tenth. He is also among the completers of a project that wins a first prize in the Science and Technology Progress Award of the Chinese Society of Image and Graphics, ranked ninth.

  7. 2025–2026

    Papers on jailbreaks and agents

    In 2025, he coauthors OptJail with Zixuan Chen, Hao Lin, Xinghao Jiang and Tanfeng Sun, a jailbreak framework for text-to-image models with safety filters. In 2026, he coauthors IterInject, a study of indirect prompt injection against LLM agents, and in October 2026 OmniGround, an omni-modal model for temporal grounding in video, audio and text.

  8. As of October 2026

    Associate researcher at Shanghai Jiao Tong

    He is an associate researcher and doctoral supervisor in the School of Computer Science (School of Cyberspace Security) at Shanghai Jiao Tong University, in the Institute of Artificial Intelligence and Cognitive Security. His research directions are video anomaly event detection, gait recognition and anonymization, and model adversarial attacks and detection.

Career

  1. NowShanghai Jiao Tong UniversityAssociate Researcher and Doctoral Supervisor

Education

  1. 2019Shanghai Jiao Tong UniversityPhD
  2. 2013Shanghai Jiao Tong UniversityBachelor's degree in engineering

From the biography.

  1. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorateIncludes Shanghai Jiao Tong University. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorate. Across 2,116 people with both an undergraduate degree and a doctorate recorded, we trace which undergraduate schools feed which doctoral programs, how often students stay at the same school, and how paths differ by birth-country region.
  2. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Shanghai Jiao Tong University. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Shanghai Jiao Tong University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Shanghai Jiao Tong University. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.

Sources and credits

Photo: We have no published photo of Ke Xu yet. Send us a free-licensed photo.

  1. 许可-上海交通大学网络空间安全学院 infosec.sjtu.edu.cn. Checked 21 October 2026.
  2. Ke Xu-上海交通大学计算机学院（网络空间安全学院、密码学院） cs.sjtu.edu.cn. Checked 21 October 2026.
  3. Gait Identification under Surveillance Environment based on Human Skeleton arxiv.org. Checked 21 October 2026.
  4. 许可-上海交通大学计算机学院（网络空间安全学院、密码学院） cs.sjtu.edu.cn. Checked 21 October 2026.
  5. About Me - Zeyu Zhao hertensu.github.io. Checked 21 October 2026.
  6. Dynamic Optimization and Safety Indicator Injection for Jailbreaking Text-to-Image Models with Multimodal Safety Filters arxiv.org. Checked 21 October 2026.
  7. IterInject: Indirect Prompt Injection Against LLM Agents via Feedback-Guided Iterative Optimization arxiv.org. Checked 21 October 2026.
  8. OmniGround: Strengthening Omni-Modal Temporal Grounding via Attention-Consistency Training arxiv.org. Checked 21 October 2026.