职称

讲师

电子邮件

chengpz@shu.edu.cn

办公室

计算机工程与科学学院403室

研究方向

多模态大模型推理，自主交互智能体，人工智能安全，网络安全

个人简介

程彭洲，博士，讲师，2026年6月毕业于上海交通大学计算机学院，师从刘功申教授和张倬胜助理教授。主要研究多模态大语言模型推理、自主交互智能体、智能体安全与网络安全。在TNNLS、TVT、ACL、ICML等重要期刊和会议发表论文26篇，其中第一或共同第一作者论文13篇，Google Scholar引用800余次；申请发明专利4项，授权3项。担任TPAMI、TDSC、TIFS等期刊及NeurIPS、ACL、ICML等会议审稿人，并担任MLNLP程序委员会主席。曾入选腾讯青云计划，蚂蚁集团蚂蚁星及小米大模型等顶尖人才计划。参与建设公益开源教程《动手学大模型》，GitHub已获4万余Star。

现招收大模型、智能体及人工智能安全方向的本科生、硕士生和科研实习生。欢迎编程基础扎实、具有科研热情和自主学习能力的同学加入，表现优秀者可推荐至头部科技企业实习。

个人主页：https://ctzhou-byte.github.io/

教育背景

2022-2026, 上海交通大学, 电子信息专业, 博士

主要工作经历

2026-至今，上海大学，讲师

2025-至今，浙江和夏科技股份有限公司，技术顾问

2024-2026，华为技术有限公司，研究实习

代表性论文或专著

[1] Dong L, Zhou Z, Cheng P, et al. Domain Adaptation of MLLM-Based GUI Agents in Documentation-Rich Environments with Standard Operating Procedures[C]//International Conference on Knowledge Science, Engineering and Management. Singapore: Springer Nature Singapore, 2026: 476-488. （CCF-C）

[2] Wu Z*, Cheng P*, Wu Z, et al. Mobile-Aptus: Confidence-Driven Proactive and Robust Interaction in MLLM-based Mobile-Using Agents[J]. IEEE Transactions on Audio, Speech and Language Processing, 2026.（共同第一作者，CCF-B，中科院二区）

[3] Hu H*, Cheng P*, Wu Z, et al. Faithful Mobile GUI Agents with Guided Advantage Estimator[C]//Forty-third International Conference on Machine Learning. （共同第一作者，CCF-A）

[4] Cheng P, Zhang X. From tool to personal assistant: The principles, evolution, and security risks of AI agents[J]. Chinese Journal of Nature, 2026, 48(2): 79-087. （第一作者，新华网全文转载）

[5] Ju T, Wang Y, Hua Y, Ma X, Cheng P, et al. Flooding spread of manipulated knowledge in llm-based multi-agent communities[J]. Science China Information Sciences, 2026, 69(7): 172103. （CCF-A，中科院一区）

[6] Wu Z, Mao R, Tian Z, Cheng P, et al. See, think, act: Teaching multimodal agents to effectively interact with gui by identifying toggles[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2026: 27536-27546. （CCF-A）

[7] Cheng P, Hu H, Wu Z, et al. Hidden Ghost Hand: Unveiling Backdoor Vulnerabilities in MLLM-Powered Mobile GUI Agents[C]// Findings of the Association for Computational Linguistics: EMNLP 2025. Suzhou, China: Association for Computational Linguistics, 2025: 7781-7805. （第一作者，CCF-B，Findings）

[8] Wu Z, Cheng P, Wu Z, et al. Gem: Gaussian embedding modeling for out-of-distribution detection in gui agents[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2026, 40(40): 33989-33997. （CCF-A）

[9] Cheng P, Liu S, Wu Z, et al. MKF-ADS: Multi-Knowledge Fusion Based Anomaly Detection System in Vehicular Control Area Networks[J]. IEEE Transactions on Vehicular Technology, 2025. （第一作者，CCF-B，中科院二区）

[10] Cheng P, Wu Z, Du W, et al. Backdoor attacks and countermeasures in natural language processing models: A comprehensive security review[J]. IEEE Transactions on Neural Networks and Learning Systems, 2025. （第一作者，CCF-B，中科院一区）

[11] Tang S*, Cheng P*, Liang H, et al. Deep‐Learning Integrated Bioelectronic‐Tissue Interface for Cardiovascular Diagnosis and Prognosis[J]. Advanced Functional Materials, 2025, 35(31): 2423264. （共同第一作者，中科院一区）

[12] Cheng P, Wu Z, Wu Z, et al. Os-kairos: Adaptive interaction for mllm-powered gui agents[C]//Findings of the Association for Computational Linguistics: ACL 2025. 2025: 6701-6725. （第一作者，CCF-A，Findings）

[13] Wu Z, Cheng P, Fang L, et al. Gracefully filtering backdoor samples for generative large language models without retraining[C]//Proceedings of the 31st International Conference on Computational Linguistics. 2025: 3267-3282. （CCF-B）

[14] Yu H, Xie T, Gui J, Wang P, Cheng P, et al. BackdoorMBTI: A Backdoor Learning Multimodal Benchmark Tool Kit for Backdoor Defense Evaluation[C]//Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V. 1. 2025: 2791-2802. （CCF-A）

[15] Cheng P, Hua L, Jiang H, et al. LSF-IDM: Deep learning-based lightweight semantic fusion intrusion detection model for automotive[J]. Peer-to-Peer Networking and Applications, 2024, 17(5): 2884-2905. （第一作者，CCF-C，中科院四区）

[16] Cheng P, Du W, Wu Z, et al. SynGhost: invisible and universal task-agnostic backdoor attack via syntactic transfer[C]//Findings of the Association for Computational Linguistics: NAACL 2025. 2025: 3530-3546. （第一作者，CCF-B，Findings）

[17] Wu Z, Zhang Z, Cheng P, et al. Acquiring clean language models from backdoor poisoned datasets by downscaling frequency space[C]//Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024: 8116-8134. （CCF-A）

[18] Li P, Cheng P, Li F, et al. Plmmark: a secure and robust black-box watermarking framework for pre-trained language models[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2023, 37(12): 14991-14999. （CCF-A）

[19] Cheng P, Han M, Liu G. DESC-IDS: Towards an efficient real-time automotive intrusion detection system based on deep evolving stream clustering[J]. Future Generation Computer Systems, 2023, 140: 266-281. （第一作者，CCF-C，中科院一区）

[20] Cheng P, Han M, Li A, et al. STC‐IDS: Spatial–temporal correlation feature analyzing based intrusion detection system for intelligent connected vehicles[J]. International Journal of Intelligent Systems, 2022, 37(11): 9532-9561. （第一作者，CCF-C，中科院一区）

[21] Cheng P, Xu K, Li S, et al. TCAN-IDS: intrusion detection system for internet of vehicle using temporal convolutional attention network[J]. Symmetry, 2022, 14(2): 310. （第一作者，中科院三区）

[22] Han M, Cheng P, Ma S. PPM-InVIDS: Privacy protection model for in-vehicle intrusion detection system based complex-valued neural network[J]. Vehicular Communications, 2021, 31: 100374. （中科院二区）

[23] Han M, Cheng P, Ma S. Cvnns-ids: Complex-valued neural network based in-vehicle intrusion detection system[C]//International Conference on Security and Privacy in Digital Economy. Singapore: Springer Singapore, 2020: 263-277.