关于我
你好！
我是中南大学计算机科学系大四本科生。
我的研究兴趣和愿景聚焦于AI Alignment、AI Safety和AI Security。随着AI系统变得越来越强大和普及，确保它们与人类价值观对齐、安全部署并抵御恶意攻击已成为该领域最关键的挑战之一。我对这一领域的承诺源于一个根本性问题：我们如何构建不仅强大，而且值得信赖、稳健且对人类有益的AI系统？
我的研究聚焦于三个相互关联的支柱：
- AI Alignment：开发方法以确保AI系统理解并遵循人类意图、价值观和伦理原则。我特别关注揭示模型如何做出决策的可解释性技术、奖励建模和偏好学习，以及用于对齐先进AI系统的可扩展监督机制。
- AI Safety：构建稳健可靠的AI系统，即使在分布偏移、对抗条件或边缘情况下也能表现出可预测的行为。这包括稳健性评估、故障模式分析和部署安全保证的研究。
- AI Security：保护AI系统免受恶意攻击，确保AI驱动应用的安全性。我的兴趣涵盖对抗稳健性、后门检测和缓解、模型提取防御以及关键系统中的安全AI部署。
我相信，推进AI Alignment、AI Safety和AI Security需要深入理解AI系统的内部机制及其现实世界的部署环境。我的方法将理论严谨性与实际影响相结合，旨在开发不仅科学可靠，而且适用于AI系统与人类和关键基础设施交互的现实场景的方法。
教育背景
新闻
- 2025.11：论文 Revisiting the Data Sampling in Multimodal Post-training from a Difficulty-Distinguish View (AAAI) 被接收！
- 2025.10：被清华大学计算机科学系录取攻读博士学位！我将专注于AI Alignment、AI Safety和AI Security研究！
- 2025.7：论文 CFSSeg: Closed-Form Solution for Class-Incremental Semantic Segmentation of 2D Images and 3D Point Clouds (ACM MM) 被接收！
- 2025.5：论文 MHALO: Evaluating MLLMs as Fine-grained Hallucination Detectors (ACL Findings) 被接收！
- 2025.4：论文 Class Incremental Semantic Segmentation Based on Linear Closed-form Solution (CVPR Workshop) 被接收！
- 2024.12：论文 AIDC: Benchmark for Analytical Learning in Incremental Disease Classification (ICASSP) 被接收！
- 2024.10：论文 HFGS: High-Frequency Information Guided Net for Multi-Regions Pseudo-CT Synthesis (BIBM) 被接收！
- 2024.9：我获得了国家奖学金（1万元）！我还获得了一等奖学金（2千元）！
论文 (完整列表)
-
-
TS-ACL: A Time Series Analytic Continual Learning Framework for Privacy-Preserving and Class-Incremental Pattern Recognition
Jiaxu Li, Fan Kejia, Songning Lai, Llinpu Lv, Anfeng Liu, Jianheng Tang, HH Song, Huiping Zhuang
审稿中
-
-
Class Incremental Semantic Segmentation Based on Linear Closed-form Solution
Jiaxu Li, Songning Lai, Rui Li, Rongchang Zhao, Yutao Yue, Huiping Zhuang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR Workshop), 2025
-
-
CFSSeg: Closed-Form Solution for Class-Incremental Semantic Segmentation of 2D Images and 3D Point Clouds
Jiaxu Li,Rui Li,Jianyu Qi,Songning Lai,Rongchang Zhao,Yutao Yue,Huiping Zhuang
ACM International Conference on Multimedia (ACM MM), 2025
-
-
AIDC: Benchmark for Analytical Learning in Incremental Disease Classification
Rongchang Zhao, Jianyu Qi, Rui Li, Zhijie Zheng,Jian Zhang,Jiaxu Li
International Conference on Acoustics, Speech, and Signal Processing(ICASSP), 2025.
-
-
ACU: Analytic Continual Unlearning for Efficient and Exact Forgetting with Privacy Preservation
jianheng Tang, Huiping Zhuang, Di Fang, Jiaxu Li, Feijiang Han, Yajiang Huang, Kejia Fan, Leye Wang, Zhanxing Zhu, Shanghang Zhang, Houbing Herbert Song, Yunhuai Liu
审稿中
-
-
CAT: Concept-level backdoor ATtacks for Concept Bottleneck Models
Songning Lai, J Yang, Y Huang, L Hu, T Xue, Z Hu, Jiaxu Li, H Liao, Yutao Yue
审稿中
-
-
HFGS: High-Frequency Information Guided Net for Multi-Regions Pseudo-CT Synthesis
Rongchang Zhao, Jianyu Qi, Rui Li, Teng Yang, Jiaxu Li, Jian Zhang, Zijian Zhang
IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2025.
-
-
MHALO: Evaluating MLLMs as Fine-grained Hallucination Detectors
Yishuo Cai, Renjie Gu, Jiaxu Li, Xuancheng Huang, Junzhe Chen, Xiaotao Gu, Minlie Huang
Association for Computational Linguistics (ACL), 2025
-
-
MR-to-CT: Only Improving Overall Similarity is Not Our Goal
Rongchang Zhao, Jianyu Qi,Jiaxu Li, Rui Li
审稿中
荣誉奖项
2024-10 国家奖学金（1万元）
2024-10 一等奖学金（2千元）