个人简介
陶仁帅,副教授,工学博士。入选北京市高层次创新创业人才支持计划、北京交通大学青年英才培育计划、2024年度中国图象图形学学会高等教育教学成果激励计划、北京市科委科技项目评审专家。主要研究方向包括计算机视觉内容理解、视觉内容取证等。主持国家自然科学基金青年项目(C类)、北京市自然科学基金重点项目,作为核心骨干(排2)参与国家自然科学基金重点项目。在ICLR、CVPR、ICCV等CCF-A/IEEE Trans国际顶级学术会议、期刊上发表论文30余篇, 担任国际顶级学术会议 ICML、ICLR 的领域主席(Area Chair)以及IJCAI 的资深程序委员(SPC)。研究成果提出了国际上首个高质量X光复杂目标检测评估基准集,收到全球200余机构的学者来信关注和使用,相关成果被广泛应用于中国海关、中国联通等实际业务场景。指导学生获北京交通大学优秀毕业论文2项。获得IJCAI 2025 论坛最佳论文奖、华为明日之星等荣誉。
报考须知:欢迎加入RST-Lab,RST是我的名字,同时也是Robust(鲁棒)、Smart(智能)、Trusted(可信)的缩写,当前人工智能还有很长的路,我们一起探索。课题组常年招收博士生(团队名额,我来带)、硕士生(本人名额),朝着发表高水平学术论文的方向努力,主页有课题组各位同学的姓名,有意加入课题组的同学可以找他们多了解。
日常思考:
1. 关于论文署名顺序的一些讨论
2. 一篇Trans期刊的研究历程分享
3. 科研入门|从论文阅读谈起
教育背景
2013-2022 北京航空航天大学计算机学院,本硕博,导师:中国科学院院士李未教授、国家杰青刘祥龙教授
工作经历
2022-2023:高级研究员,华为诺亚方舟实验室
2023-至今:副教授,北京交通大学计算机学院
关于招生
与其浑浑噩噩度日,何不与我共图大事?
RST-Lab是个年轻的课题组,我和学生一起朝着以发表高水平研究论文的方向努力。我将为大家的发展提供有竞争力的专业指导(组会、论文分享、代码指导)、未来发展的深造/工作资源(大牛推荐信、深造机会、工作机会)和学习条件(助研补贴、算力、机器屏幕、参加学术会议资助)。欢迎热爱学术研究、将来打算深造或从事专业技术工作的有志青年联系了解。
算力情况:
目前课题组拥有8张H200显卡(显存140G,可搞大模型)、22张4090显卡(8张高配版(显存48G)+14张标准版(显存24G)),4张3090显卡,若干3080显卡。另根据每位同学的具体研究任务需要,可协调企业更高算力配置。总体来说,算力资源较为充足,不会让任何一位同学因算力瓶颈而阻碍研究。
工作条件:
已入学研究生都已配备学校工位、台式机电脑、4K高清大屏幕(RMB 1000+)、定制版键盘鼠标。另外,每位同学在科研上的支出,实报实销。
团队氛围:
团队定期举行聚餐活动。
同时鼓励大家多组织运动活动,积极锻炼,健康生活,在学校运动会上取得成绩的同学会有额外奖励。
投稿论文、发表论文的时候会有奖励。
资助发表论文的同学参加学术会议进行交流。
一、攻读博士、硕士学位学生:
招生专业:
计算机科学与技术、控制科学与工程、计算机技术、人工智能、软件工程、新一代电子信息技术(含量子技术等)
二、本科生:
(表现优秀的本科生如愿意留在课题组读研则优先考虑,不愿意则可协助推荐其他去处,来去自由)
申请目的包括但不限于:
-
积累研究经历、学术成果,为将来继续深造(攻读硕士、博士学位)打下基础
-
完成毕业设计、参加竞赛、大创等活动提高能力
-
需要实习证明、推荐信等个人发展所需材料
申请条件:
-
时间上能保证6个月以上
-
对深度学习有一定了解,掌握基本的深度学习框架(PyTorch)
-
对发表顶级期刊、会议论文有自驱力
论文/期刊
论文情况更新至2025.06,后续可移步至谷歌学术 https://scholar.google.com/citations?user=rAGw-fcAAAAJ&hl=zh-CN 查看详细 List
Under Review
-
Sh. Yao, R. Tao*, Ch. Zhang, X. Zheng, Y. Zhao. ForGeNet: Towards Real-World Deepfake Detection under Imbalanced Training Data Scenarios. (Under Review)
-
Ch. Peng, R. Tao*, Zh. Ren, X. Liu, Y. Wei. Can a Second-View Image Be a Language? Geometric and Semantic Cross-Modal Reasoning for X-ray Prohibited Item Detection. (Under Review)
-
G. Bao, R. Tao*, J. Yang, G. Luo, H. Bai. Eliminating Subtle Shifts for Real-World Industrial Domain Adaptation. (Under Review)
-
Zh. Yang, R. Tao*, Ch. Zhang, Zh. Liu, X. Zheng, Y. Zhao. Asymmetric Anchoring: Opening the Black Box of MLLMs for Forgery Detection. (Under Review)
-
Zh. Yang, R. Tao*, X. Zheng, G. Yang, Ch. Zhang. Leveraging Unlabeled Data from Unknown Sources via Dual-Path Guidance for Deepfake Face Detection. (Under Review)
-
Z. Qin (本科生), Y. Ji, R. Tao, Y. Tian, Y. Liu, Y. Wang, X. Zheng. Scaling Up AI-Generated Image Detection with Generator-Aware Prototypes. (Under Review)
-
M. Li (本科生), R. Tao*, Y. Liu, Ch. Tan, H. Qin, B. Li, Y. Wei, Y. Zhao. Pay Less Attention to Deceptive Artifacts: Robust Detection of Compressed Deepfakes on Online Social Networks. (submitted to IJCV, Under Review)
-
J. Wang, P. Zhang (本科生), R. Tao*, J. Yang, H. Liu, X. Liu, Y. Wei, Y. Zhao. Don't Modify without Permission: Triggering Backdoors via Model Operation. (Under Review)
-
H. Wang, R. Tao, W. Wang, Y. Wei. PAD-F: Prior-Aware Debiasing Framework for Long-Tailed X-ray Prohibited Item Detection. (Under Review)
-
W. Liu, R. Tao, H. Zhu, Y. Sun, Y. Zhao, Y. Wei. BGM: Background Mixup for Prohibited Item Detection. (Under Review)
-
J. Sun, H. Zhu, W. Liu, Y. Sun, R. Tao, Y. Wei. Taming Generative Synthetic Data for X-ray Prohibited Item Detection. (Under Review)
Publication
-
J. Wang, H. Liu, R. Tao*, J. Sun, X. Liu, Y. Zhao. Practical and Flexible Backdoor Attack against Deep Learning Models via Shell Code Injection. IEEE TIFS. (CCF-A)
-
Ch. Tan, X. Ming, J. Wang, R. Tao, B. Li, Y. Wei, Y. Zhao, Y. Lu. Semantic Visual Anomaly Detection and Reasoning in AI-Generated Images. ICLR 2026.
-
W. Nie (本科生), Z. Li, R. Tao*, B. Wu, Y. Wei, Y. Zhao. CoCoDiff: Correspondence-Consistent Diffusion Model for Fine-grained Style Transfer. ICLR 2026.
-
Sh. Yao, R. Tao*, X. Zheng, C. Liang, Ch. Zhang. Leveraging Failed Samples: A Few-Shot and Training-Free Framework for Generalized Deepfake Detection. AAAI 2026. (CCF-A)
-
R. Li (本科生), J. Wang, H. Chen, H. Ding, J. Zhou, R. Tao*. Dormant Backdoor: Weaponizing Model Finetuning for Feasible Backdoor Attacks Against Pretrained Models. AAAI 2026. (CCF-A)
-
H. Chen (本科生), R. Li (本科生), Y. Guo (本科生), R. Tao*, H. Bai, Y. Zhao. Dialectical Chain Distillation: Transferring Dialectical Reasoning from Teacher–Student Interactions to Small Language Models. Best paper award, IJCAI Workshop 2025.
-
R. Tao, Ch. Tan, H. Liu, J. Wang, H. Qin, Y. Chang, W. Wang, R. Ni, Y. Zhao. SAGNet: Decoupling Semantic-Agnostic Artifacts from Limited Training Data for Robust Generalization in Deepfake Detection. IEEE TIFS 2025. (CCF-A) 研究历程分享见日常思考链接
-
Zh. Yang, R. Tao*, Ch. Zhang, X. Zheng. Multi-Cue Fusion for Forgery Detection: A Spatial-Frequency and Chromatic Approach. IEEE ISI 2025, oral.
-
R. Tao, H. Chen (本科生), Y. Guo, J. Wang, B. Wang, R. Ni, Y. Zhao. LS-PRISM: A Layer-Selective Pruning Method via Low-Rank Approximation and Sparsification for Efficient Large Language Model Compression. Neural Network. (SCI Q1)
-
R. Tao, M. Li (本科生), Ch. Tan, H. Liu, H. Qin, Y. Zhao. ODDN: Addressing Unpaired Data Challenges in Open-World Deepfake Detection on Online Social Networks. AAAI 2025, oral, ~top 4.6%. (CCF-A)
-
Ch. Tan, R. Tao, G. Gu, B. Wu, Y. Zhao, Y. Wei. C2P-CLIP: Injecting Category Common Prompt in CLIP to Enhance Generalization in Deepfake Detection. AAAI 2025. (CCF-A)
-
W. Feng, H. Qin, Ch. Yang, Zh. An, L. Huang, B. Diao, F. Wang, R. Tao, Y. Xu, M. Magno. MPQ-DM: Mixed Precision Quantization for Extremely Low Bit Diffusion Models. AAAI 2025, oral, ~top 4.6%. (CCF-A)
-
R. Tao, Z. Qin (本科生), Y. Ding, Ch. Tan, J. Wang, W. Wang. Unlocking the Potential of Lightweight Quantized Models for Deepfake Detection. IJCAI 2025. (CCF-A)
-
R. Tao, Sh. Tang (本科生), H. Qin, W. Wang, Y. Wei, Y. Zhao. LEDNet: A Multimodal Foundation Model for Robust Deepfake Detection. SCIENCE CHINA Information Science 2025. (CCF-A)
-
R. Tao, H. Wang, Y. Guo (本科生), H. Chen (本科生), L. Zhang, X. Liu, Y. Wei, Y. Zhao. Dual-view X-ray Detection: Can AI Detect Prohibited Items from Dual-view X-ray Images like Humans? CVPR 2025. (CCF-A, 博士时期研究工作的延续)
-
L. He, Y. Chang, R. Cong, H. Liu, Sh. Huang, R. Tao, Y. Zhao. Rethinking Depth Guided Reflection Removal. IEEE TMM, 2024. (SCI Q1)
-
Y. Xue, H. Hao, J. Wang, R. Tao, Y. Liang, P. Feng, X. Liu. Vision-fused Attack: Advancing Aggressive and Stealthy Adversarial Text against Neural Machine Translation. IJCAI 2024. (CCF-A)
-
R. Tao, H. Li, T. Wang, Y. Wei, Y. Ding, B. Jin, H. Zhi, X. Liu, A. Liu. Exploring Endogenous Shift for Cross-domain Detection: A Large-scale Benchmark and Perturbation Suppression Network. CVPR, 2022. (CCF-A)
-
R. Tao, Y. Wei, X. Jiang, H. Li, H. Qin, J. Wang, Y. Ma, L. Zhang, X. Liu. Towards Real-world X-ray Security Inspection: A High-Quality Benchmark And Lateral Inhibition Module For Prohibited Items Detection. ICCV, 2021. (CCF-A)
-
R. Tao, T. Wang, Z. Wu, C. Liu, A. Liu, X. Liu. Few-shot X-ray Prohibited Item Detection: A Benchmark and Weak-feature Enhancement Network. ACM MM, 2022. (CCF-A)
-
Y. Wu, J. Fan, R. Tao*, J. Wang, H. Qin, A. Liu, X. Liu. Sequential alignment attention model for scene text recognition. JVCI 2021. (SCI Q2)
-
Y. Wei, R. Tao, Zh. Wu, Y. Ma, L. Zhang, X. Liu. Occluded Prohibited Items Detection: An X-ray Security Inspection Benchmark and De-occlusion Attention Module. ACM MM 2020, oral,(CCF-A)