Gaojie (Jay) Jin

Gaojie (Jay) Jin

I am an Assistant Professor in AI & CS at the University of Macau, previously I worked as an Assistant Professor at the University of Exeter, UK, and as a Research Associate Professor at the Chinese Academy of Sciences.

I am interested in developing reliable (e.g., robust, generalisable, and efficient [NeurIPS20,CVPR22,23,ICCV23a,TIFS25,ICLR25a]) algorithms for modern machine learning models and applications [ICCV23b,AAAI23,ECCV24]. I focuse on aligning AI with human rules/preferences [ICML24,ICLR25b], with an emphasis on providing theoretical guarantees and statistical analyses [TMLR22,TPAMI25].

Currently, my research interests include, but are not limited to:

  • Trustworthy Human–GenAI Alignment: Developing unified statistical frameworks, e.g., combining PAC-Bayesian theory, conformal prediction, and adversarial analysis, to provide rigorous guarantees for human–GenAI alignment. [ICML26a,EMNLP26]

  • Efficiency Robustness of GenAI: Building empirical and theoretical foundations to analyse and mitigate vulnerabilities in reasoning efficiency. [ICML26b,AAAI26]

I am pleased to announce multiple openings for PhD and post-doctoral positions for the 2026/27 academic year. If you are interested in these opportunities or would like to discuss potential collaborations, please don’t hesitate to contact me at gaojiejin at um dot edu dot mo.

Recent News

  • (08/2026) I will be serving as an Area Chair for ICLR 2027.
  • (08/2026) One paper accepted to EMNLP 2026 (Findings), one paper accepted to MICCAI 2026.
  • (05/2026) Served as a PhD Viva external examiner for a candidate at King’s College London.
  • (04/2026) Three papers accepted to ICML 2026, congrats to all coauthors, congrats to Xinyu for his first paper.
  • (03/2026) I will be serving as an Area Chair for NeurIPS 2026.
  • (02/2026) One paper accepted to CVPR 2026 (Oral), one paper accepted to IEEE ISIT 2026.
  • (01/2026) One paper accepted to ICLR 2026, one paper accepted to IEEE TIP, one paper accepted to ICASSP 2026.
  • (11/2025) I will be serving as an Associate Editor for Theoretical Computer Science.
  • (11/2025) Two papers accepted to AAAI 2026.
  • (10/2025) Got a grant from Isambard-AI with 10000 GPU hours.
  • (09/2025) Got a grant from NVIDIA Academic Grant Program.
  • (09/2025) I will be serving as an Area Chair for ICLR 2026.
  • (06/2025) One paper accepted to TMLR.
  • (06/2025) One paper accepted to IEEE TPAMI.
  • (05/2025) I will be serving as an Area Chair for NeurIPS 2025 (Position Paper Track).
  • (04/2025) One paper accepted to IEEE TIFS.
  • (02/2025) Got two PhD studentships from the EU Horizon project.
  • (01/2025) Two papers accepted to ICLR 2025.
  • (07/2024) Start my position as a Lecturer (Assistant Professor) at Exeter.
  • (06/2024) One paper accepted to ECCV 2024.
  • (01/2024) One paper accepted to ICML 2024.

Funding & Grants

  • Exploiting Robustness of Reasoning Efficiency in Agentic AI (PI) (2025-2026)
    Funded by AIRR Isambard-AI (10000 GPU hours)
  • Exploiting Overthinking Attacks on GenAI (PI) (2025-2026)
    Funded by NVIDIA Academic Grant Program (NVIDIA DGX Spark)
  • Robustifying Generative AI through Human-Centric Integration of Neural and Symbolic Methods (External Participant) 2025 - 2028
    Funded by EU Horizon.
  • FOCETA (Foundations for Continuous Engineering of Trustworthy Autonomy) (Research Assistant) 2021 - 2023
    Funded by EU H2020.
  • EnnCore (End-to-End Conceptual Guarding of Neural Architectures) (Research Assistant) 2020 - 2024
    Funded by EPSRC.
  • SOLITUDE (Safety Argument for Learning-enabled Autonomous Underwater Vehicles) (Research Assistant) 2020 - 2022
    Funded by UK DSTL.

Publications

#:Equal Contribution, ✉️:Corresponding Author

2026

  1. GradientStabilizer: Fix the Norm, Not the Gradient
    Tianjin Huang, Zhangyang Wang, Haotian Hu, Zhenyu Zhang, Gaojie Jin, et al.
    International Conference on Machine Learning (ICML), 2026.

  2. Margin-Adaptive Confidence Ranking for Reliable LLM Judgement
    G. Jin, Y. Tao, L. Yu, T. Huang.
    International Conference on Machine Learning (ICML), 2026.

  3. OTora: A Unified Red Teaming Framework for Reasoning-Level Denial-of-Service in LLM Agents
    X. Li, R. Mu, L. Li, T. Huang, G. Jin✉️.
    International Conference on Machine Learning (ICML), 2026.

  4. Preference Alignment on Diffusion Models: A Comprehensive Survey for Image Generation and Editing
    S. Wu, X. Si, C. Xing, J. Wang, G. Jin, G. Cheng, X. Huang.
    Computer Science Review, 61, 100900, 2026.

  5. BadThink: Triggered Overthinking Attacks on Chain-of-Thought Reasoning in Large Language Models
    S. Liu, R. Li, L. Yu, L. Zhang, Z. Liu, G. Jin✉️.
    AAAI Conference on Artificial Intelligence (AAAI), 2026.

  6. CluCERT: Certifying LLM Robustness via Clustering-Guided Denoising Smoothing
    Z. Wang, G. Jin, J. Hu, R. Mu.
    AAAI Conference on Artificial Intelligence (AAAI), 2026.

  7. Confusion-Aware Spectral Regularizer for Long-Tailed Recognition
    Z. Zhu#, G. Jin#, H. Zhu#, S. Y. Lu#, Y. Zhang, Z. Fu, R. Mu, G. Zhang, Z. Sun, et al.
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026. Oral.

  8. Dual-Kernel Adapter: Expanding Spatial Horizons for Data-Constrained Medical Image Analysis
    Z. Zhu, H. Zhu, S. Lu, X. Li, Y. Meng, G. Jin, L. Yin, L. Hu, D. Wang, L. Liu, et al.
    International Conference on Learning Representations (ICLR), 2026.

  9. StealthMark: Harmless and Stealthy Ownership Verification for Medical Segmentation via Uncertainty-Guided Backdoors
    Q. Yu, C. Zhang, G. Jin, T. Huang, W. Zhou, W. Li, X. Jin, B. Huang, Y. Zhao, et al.
    IEEE Transactions on Image Processing (TIP), 2026.

  10. Localize-Then-Decide Guarantees for LLM Judgments
    X. Li, Y. Zhou, G. Cao, Z. Fu, T. Huang, G. Jin✉️.
    Findings of the Association for Computational Linguistics: EMNLP, 2026.

  11. CPR: Chained Perceptual Refinement for Coarse-to-Fine Medical Image Classification
    S. Y. Lu, H. Zhu, Z. Zhu, G. Jin, Z. Fu, L. Yin, K. Li, L. Liu, T. Huang.
    International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2026.

  12. A Unified Framework for PAC-Bayesian Norm-based Generalization Bounds
    X. Yi, G. Jin, X. Huang, S. Jin.
    IEEE International Symposium on Information Theory (ISIT), 2026, pp. 1–6.

  13. TRAJRS: Towards Certified Robustness in Pedestrian Trajectory Prediction
    L. Zhang, G. Jin, Y. Shi, Q. Li, C. C. Huang, D. N. Jansen, L. Zhang.
    IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026.

2025

  1. Safeguarding Large Language Models: A Survey
    Y. Dong, R. Mu, Y. Zhang, S. Sun, T. Zhang, C. Wu, G. Jin, Y. Qi, J. Hu, J. Meng, et al.
    Artificial Intelligence Review, 58(12), 382, 2025.

  2. SPAM: Spike-Aware Adam with Momentum Reset for Stable LLM Training
    T. Huang, Z. Zhu, G. Jin, L. Liu, Z. Wang, S. Liu.
    International Conference on Learning Representations (ICLR), 2025.

  3. Enhancing Robust Fairness via Confusional Spectral Regularization
    G. Jin, S. Wu, J. Liu, T. Huang, R. Mu.
    International Conference on Learning Representations (ICLR), 2025.

  4. Invariant Correlation of Representation with Label: Enhancing Domain Generalization in Noisy Environments
    G. Jin, R. Mu, X. Yi, X. Huang, L. Zhang.
    IEEE Transactions on Information Forensics and Security (TIFS), 2025.

  5. Toward Linearly Regularizing the Geometric Bottleneck of Linear Generalized Attention
    J. Liu, X. Yi, X. Yin, Y. Song, G. Jin, X. Huang.
    Transactions on Machine Learning Research (TMLR), 2025.

  6. S²O: Enhancing Adversarial Training with Second-Order Statistics of Weights
    G. Jin, X. Yi, W. Huang, S. Schewe, X. Huang.
    IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2025.

2024

  1. Position: Building Guardrails for Large Language Models Requires Systematic Design
    D. Yi, R. Mu, G. Jin, Y. Qi, J. Hu, X. Zhao, J. Meng, W. Ruan, X. Huang.
    International Conference on Machine Learning (ICML), 2024.

  2. Formal Verification of Robustness and Resilience of Learning-Enabled State Estimation Systems
    W. Huang, Y. Zhou, G. Jin, Y. Sun, J. Meng, F. Zhang, X. Huang.
    Neurocomputing, 585, 127643, 2024.

  3. Class-Aware Cross Pseudo Supervision Framework for Semi-Supervised Multi-organ Segmentation in Abdominal CT Scans
    D. Yang, H. Zhao, G. Jin, H. Meng, L. Zhang.
    Chinese Conference on Pattern Recognition and Computer Vision (PRCV), 2024.

  4. Out-of-Bounding-Box Triggers: A Stealthy Approach to Cheat Object Detectors
    T. Lin, L. Yu, G. Jin, R. Li, P. Wu, L. Zhang.
    European Conference on Computer Vision (ECCV), 2024.

2023

  1. A Survey of Safety and Trustworthiness of Large Language Models through the Lens of Verification and Validation
    X. Huang, W. Ruan, W. Huang, G. Jin, Y. Dong, C. Wu, S. Bensalem, R. Mu, et al.
    Artificial Intelligence Review, 2023.

  2. Randomized Adversarial Training via Taylor Expansion
    G. Jin, X. Yi, D. Wu, R. Mu, X. Huang.
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023.

  3. SAFARI: Versatile and Efficient Evaluations for Robustness of Interpretability
    W. Huang, X. Zhao, G. Jin, X. Huang.
    IEEE/CVF International Conference on Computer Vision (ICCV), 2023.

  4. Optimising Event-Driven Spiking Neural Network with Regularisation and Cutoff
    D. Wu, G. Jin, H. Yu, X. Yi, X. Huang.
    Frontiers in Neuroscience, 2023.

  5. Certified Policy Smoothing for Cooperative Multi-Agent Reinforcement Learning
    R. Mu, W. Ruan, L. S. Marcolino, G. Jin, Q. Ni.
    AAAI Conference on Artificial Intelligence (AAAI), 2023.

  6. TrajPAC: Towards Robustness Verification of Pedestrian Trajectory Prediction Models
    L. Zhang, N. Xu, P. Yang, G. Jin✉️, C. C. Huang, L. Zhang.
    IEEE/CVF International Conference on Computer Vision (ICCV), 2023.

  7. Machine Learning Safety
    X. Huang, G. Jin, W. Ruan.
    Machine Learning Safety, pp. 3–13, 2023.

2022

  1. S²O: Enhancing Adversarial Training with Second-Order Statistics of Weights
    G. Jin, X. Yi, W. Huang, S. Schewe, X. Huang.
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
    (Later extended in IEEE TPAMI, 2025.)

  2. Weight Expansion: A New Perspective on Dropout and Generalization
    G. Jin, X. Yi, P. Yang, L. Zhang, S. Schewe, X. Huang.
    Transactions on Machine Learning Research (TMLR), 2022.

2020

  1. How Does Weight Correlation Affect the Generalisation Ability of Deep Neural Networks
    G. Jin, X. Yi, L. Zhang, L. Zhang, S. Schewe, X. Huang.
    Advances in Neural Information Processing Systems (NeurIPS), 2020.

Full list

Teaching

  • ECM1416: Computational Mathematics
  • COMM113: Deep Learning

Supervised Students (as Primary Supervisor)

  • Xinyu Li (PhD candidate at Exeter since 02/2026)
    Published: ICML 2026, EMNLP 2026 (Findings)
  • Qiutong Xu (PhD candidate at Exeter, expected since 11/2026)
  • Jingxiao Li (PhD candidate at Macau since 09/2026)
  • Zekang Wang (PhD candidate at Macau since 09/2026)
  • Hongyi Zhang (Master Student at Macau since 09/2026)
  • Yuchen Liu (Master Student at Macau since 09/2026)

Academic Service

  • Reviewer
    ICML, NeurIPS, ICLR, AISTATS, AAAI, CVPR, ICCV, ECCV, JMLR, IJCV, TMLR, TIFS, TDSC
  • Area Chair
    NeurIPS 2025/2026, ICLR 2026/2027
  • Associate Editor
    Theoretical Computer Science
  • Organise the workshop TrustRL: Trustworthy in Reinforcement Learning at ATC 2024