Lei Li

Research Interests

My research focuses on federated learning, efficient AI systems, and mathematical optimization, with emphasis on:

• Federated LearningEfficient sparse and multi-task learning across distributed clients, with an emphasis on adaptive training and communication efficiency.
• Efficient AI SystemsFast, accuracy-aware compression for large language models and tensor-native GPU codecs that reduce latency and energy use.
• Mathematical OptimizationOptimization methods for dynamic sparsity, min-max learning, and robust multi-task neural networks.

Publications

1. NOVA: A Tensor-Native GPU Codec for Fast and Accuracy-Aware LLM Weight Compression

   Lei Li, Haochen Yang, Ben Mechels, Tianyun Zhang, Caiwen Ding

   HPCA2027

   01Under reviewFirst author

2. DyFreeze: A Dynamic Parameter-Freezing Framework for Federated Sparse Training

   Lei Li, Haochen Yang, Jiacheng Guo, Hongkai Yu, Minghai Qin, Tianyun Zhang

   AAAI2027

   02Under reviewFirst author

3. Adaptive Parameter Freezing and Sparse Upload for Efficient Federated Dynamic Sparse Training

   Lei Li, Haochen Yang, Jiacheng Guo, Hongkai Yu, Minghai Qin, Tianyun Zhang

   Pattern Recognition2027

   03Under reviewFirst author

4. FedSTA: Spatio-Temporal Alternation for Efficient Federated Multi-Task Learning

   Lei Li, Haochen Yang, Jiacheng Guo, Hongkai Yu, Minghai Qin, Tianyun Zhang

   TCSVT2026

   04JournalFirst author

5. An Efficient and Accurate Dynamic Sparse Training Framework Based on Parameter Freezing

   Lei Li, Haochen Yang, Jiacheng Guo, Hongkai Yu, Minghai Qin, Tianyun Zhang

   AAAI2025

   05ConferenceFirst author

6. Task-Aware Federated Multi-Task Learning

   Lei Li, Haochen Yang, Jiacheng Guo, Hongkai Yu, Minghai Qin, Tianyun Zhang

   MMAsia2025

   06ConferenceFirst author

7. Unsupervised Multi-Agent and Single-Agent Perception from Cooperative Views

   Haochen Yang, Baolu Li, Lei Li, Delin Ren, Jiacheng Guo, Minghai Qin, Tianyun Zhang, Hongkai Yu

   CVPR2026

   07Conference

8. DA3D: Domain-Aware Dynamic Adaptation for All-Weather Multimodal 3D Detection

   Haochen Yang, Lei Li, Jiacheng Guo, Baolu Li, Minghai Qin, Hongkai Yu, Tianyun Zhang

   ACM MM2025

   08Co-first author

9. Robust Multi-Task Adversarial Attacks Using Min-Max Optimization

   Jiacheng Guo, Lei Li, Haochen Yang, Baocheng Geng, Hongkai Yu, Minghai Qin, Tianyun Zhang

   ICASSP2025

   09Conference

10. A Min-Max Optimization Framework for Sparse Multi-Task Deep Neural Network

    Jiacheng Guo, Lei Li, Huiming Sun, Minghai Qin, Hongkai Yu, Tianyun Zhang

    Neurocomputing2025

    10Journal

11. Stealthy Multi-Task Adversarial Attacks

    Jiacheng Guo, Tianyun Zhang, Lei Li, Haochen Yang, Hongkai Yu, Minghai Qin

    ECCV2026

    11Conference

Background