智慧在前进✊

教育经历

• 2018-09至2024-07 中国科学院信息工程研究所（物联网安全信息技术北京市重点实验室），网络空间安全，博士
• 2014-09至2018-07 大连理工大学，信息与计算科学，学士

研究方向

面向具身无人系统与自主智能体，研究复杂对抗环境下的行为一致性安全问题，重点关注系统在感知、决策、协同与执行过程中，其实际运行行为与预期目标之间的偏离、篡改与失控风险，并探索运行时验证、可信协同与自主安全演化机制。

• 运行时行为验证：

  面向边缘计算、具身无人系统与协同智能体场景，研究第三方节点与自主智能体的运行时行为验证、结果正确性检测与异常行为溯源方法，保障系统实际执行行为与预期目标之间的一致性与可信性。

• 自主智能体协同与对抗安全：

  面向复杂对抗环境，研究基于大语言模型与自主智能体的协同攻防机制，重点关注多智能体协同、自主侦察、动态任务规划与长时程行为演化中的安全性、可信性与鲁棒性问题。

• 供应链安全：

  面向复杂软件供应链与第三方服务生态，研究软硬件组件、开源依赖、第三方模型与云边服务中的可信溯源、风险传播与运行时验证问题，探索供应链攻击的检测、分析与动态防护机制。

主要学术成果

研究项目

• [1] 具身智能体的任务意图理解与安全执行机制研究，互联网体系结构全国重点实验室开放课题，2025.12-2027.12，主持.
• [2] XX要素模型相关算法（4），北方自动控制技术研究所，2025.01-2025.12，主持.
• 曾参与国家重点研发计划项目、国家自然科学基金重点项目、北京市科技计划项目、信息工程研究所国际合作项目等多项科研项目或课题。

学术成果

• [1] Zhihui Zhao, Xiaorong Dong, Yaowen Zheng, Xiaohui Chen, Yimo Ren, Hangbei Cheng, Yongle Chen, and Limin Sun. Breaking Cross-modal Alignment in Embodied Intelligence: A Multimodal Adversarial Attack Framework for Vision-Language-Action Models. In Proceedings of the ACM Web Conference 2026 (WWW 26).(CCF A)
• [2] Zhihui Zhao, Xiaorong Dong, Yimo Ren, Jianhua Wang, Dan Yu, Hongsong Zhu, Yongle Chen. SMTFL: Secure Model Training to Untrusted Participants in Federated Learning. IEEE Transactions on Mobile Computing (TMC). 2026, doi: 10.1109/TMC.2026.3664822. (CCF A)
• [3] Zhihui Zhao, Haoyu Bin, Hong Li, Nan Yu, Hongsong Zhu, Limin Sun. FeaShare: Feature Sharing for Computation Correctness in Edge Preprocessing[J]. IEEE Transactions on Mobile Computing, 2024. doi: 10.1109/TMC.2024.3391294. (CCF A)
• [4] Zhihui Zhao, Yicheng Zeng, Jinfa Wang, Hong Li, Hongsong Zhu, and Limin Sun.Detection and Incentive: A Tampering Detection Mechanism for Object Detection in Edge Computing[C]. IEEE 41st International Symposium on Reliable Distributed Systems (SRDS). IEEE, 2022: 166-177.Vienna, Austria, September 19-22, 2022. (CCF B)
• [5] Zhihui Zhao, Yuan Jin, Siyan Zhu, Dan Yu, Hongsong Zhu, Yongle Chen, RDGV: Reputation-driven Gradual Verification for Tampering Localization in Cooperative Task Offloading, IEEE Transactions on Reliability, 2025. (中科院2区)
• [6] Zhihui Zhao, Sen Zhao, Fei Lv, Shuaizong Si, Hongsong Zhu, and Limin Sun. RIETD: A Reputation Incentive Scheme Facilitates Personalized Edge Tampering Detection[J]. IEEE Internet of Things Journal,vol.11,no.8, pp.14771-14788. (中科院2区Top)
• [7] Zhihui Zhao, Weizhong Wang, Hongsong Zhu, Hong Li, Limin Sun, Sen Zhao,Yan Hu.EdgeCC: An Authentication Framework for the Fast Migration of Edge Services Under Mobile Clients[C].Wireless Algorithms, Systems, and Applications: 15th International Conference (WASA) 2020, Proceedings,PartI 2020: 755-767.Qingdao, China, September 13-15, 2020.(CCF C)
• [8] Xiaorong Dong, Zhihui Zhao*, Hangbei Cheng, Yimo Ren, Dan Yu, Yongle Chen. MTSec: AIGC-Enhanced Security Model Training for Multimodal Federated Learning[J]. Knowledge-Based Systems, 2025: 114748. (中科院1区Top, 指导学生一作)
• [9] Dan Yu, Juntao Shi, Jiaqian Ren, Zhaoteng Yan, Zhihui Zhao*, Yongle Chen, Enhancing Security in Embodied Intelligence: Attack Detection via Constraint Functions,International Conference on Algorithms and Architectures for Parallel Processing (ICA3PP), 2025 (CCF C)
• [10] Yongpan Wang*, Qi Liu*, Zhen Lei, Siyuan Li, Xiaojie Zhu, Zhihui Zhao, Xiaodong Gu, Yongle Chen. BINENHANCE-PRO: Enhancing Binary Code Search by Distinguishing Similar but Non-Homologous Functions, IEEE Transactions on Dependable and Secure Computing (TDSC), 2026. (CCF A, 指导学生共同一作)
• [11] 赵智慧. 边缘篡改检测与定位:现状、挑战与应对，西安交通大学出版社，2025.
• [12] Yimo Ren, Jinfa Wang, Zhihui Zhao, Hui Wen, Hong Li, Hongsong Zhu, Automated tactics Planning for Cyber Attack and Defense based on Large Language Model Agents, Neural Networks, 2025:191, 107842. (CCF B)
• [13] Sen Zhao, Jinfa Wang, Shouguo Yang, Yicheng Zeng, Zhihui Zhao, Hongsong Zhu*, Limin Sun. ProsegDL: Binary Protocol Format Extraction by Deep Learning-based Field Boundary Identification[C]. IEEE 30th International Conference on Network Protocols (ICNP), 2022, 1-12. (CCF B)
• [14] Ying Liang, Zhihui Zhao, Yaowen Zheng, Zhen Dai, Yongle Chen, Limin Sun. RoboClarify: Clarifying Ambiguous Instructions through Scenario-Guided Risk Assessment for Home Embodied Agents, International Conference on Algorithms and Architectures for Parallel Processing (ICA3PP), 2025. (CCF C, 指导学生一作)
• [15] Jiaxin Gao, Yaohua Liu, Danchen Cui, Zhihui Zhao, SNOL: Subtle Nested Objective Learning for Enhancing and Super-Resolving Ultra-Low-Light Images， ACM International Conference on Multimedia Retrieval (ICMR) 2026. (CCF B)
• [16] Yuhang Lu, Zhen Wang, Sen Zhao, Zhihui Zhao, Yongji Liu, Yimo Ren, Hongsong Zhu, Synergistic Struture-aware Guided Network for Binary Protocol Format Inference, ICASSP 2026. (CCF B)
• [17] Yicheng Zeng. Jiaqian Peng, Zhihui Zhao, Zhanwei Song, Hongsong Zhu, Limin Sun, SIFOL: Solving Implicit Flows in Loops for Concolic Execution, IEEE International Performance, Computing, and Communications Conference (IPCCC), Austin, TX, USA, 2022, pp. 290-297. (CCF C)

国家发明专利

• [1] 朱红松，赵智慧，李志，李红，于楠，孙利民. 一种边缘服务迁移方法及装置，2023-05-30，授权号：CN 111381962 B
• [2] 于楠，赵智慧，朱红松，王进法，孙利民. 数据传输与处理可信认证的方法、装置及电子设备，2024-05-26，授权号：CN 114520733 B
• [3] 于丹，金源，赵智慧，陈永乐，王建华.一种适用于任务卸载场景的边缘节点异常行为检测与定位方法，授权号：CN 119696915 B

标准

• [1] 中国人工智能产业发展联盟，具身智能总体架构及技术能力要求，AIIA/T 0198-2024，2025-01-10.(已发布，参与)
• [2] 中国人工智能产业发展联盟，具身智能产品 能力要求与评估方法 第1部分：家庭陪伴机器人，AIIA/T 0199-2024，2025-01-10.(已发布，参与)
• [3] 中国网络安全产业联盟，网络安全运营大模型参考架构，2026-03-16.（已发布，参与）
• [4] 中国信息协会，人工智能大模型生成内容安全评测方法，2026-01-16.（公开征求意见）

社会兼职与荣誉

• [1] CCF网络与系统安全专委会执行委员
• [2] 《Digital Twins and Applications》期刊青年编委
• [3] ACM太原分会新星奖（2024年度）
• [4] 广东省数字安全协会/数字政府网络安全产业联盟特聘专家
• [5] 中国物联网学术会议（CWSN 2024）组织委员会成员
• [6] CyberC 2025会议大数据竞赛主席
• [7] 曾担任TIFS、TMC、AAAI、IoTJ、Cybersecurity、Globecom、IPCCC、ICA3PP、PRCV、计算机研究与发展、信息安全学报、信息网络安全等国内外期刊或会议审稿人