publications
publications by categories in reversed chronological order. generated by jekyll-scholar.
2026
- Mixture of Distributions Matters: Dynamic Sparse Attention for Efficient Video Diffusion TransformersIn International Conference on Machine Learning (ICML), 2026
- CR-Net: Scaling Parameter-Efficient Training with Cross-Layer Low-Rank StructureIn International Conference on Learning Representations (ICLR), 2026
2025
- MISA: Memory-Efficient LLMs Optimization with Module-wise Importance SamplingIn The Conference on Neural Information Processing Systems (NeurIPS), 2025