Publications

• Accelerating Transformer Pre-training with 2:4 Sparsity [arXiv] [OpenReview] [PDF] [Project page]

  Yuezhou Hu, Kang Zhao, Weiyu Huang, Jianfei Chen, Jun Zhu

  International Conference on Machine Learning (ICML), 2024

• S-STE: Continuous Pruning Function for Efficient 2:4 Sparse Pre-training [arXiv] [OpenReview] [Project page]

  Yuezhou Hu, Jun Zhu, Jianfei Chen

  Neural Information Processing Systems (NeurIPS), 2024

• Pruning Large Language Models with Semi-Structural Adaptive Sparse Training [arXiv] [Project page]

  Weiyu Huang, Yuezhou Hu, Guohao Jian, Jun Zhu, Jianfei Chen

  AAAI Conference on Artificial Intelligence, 2025

• AdaSPEC: Selective Knowledge Distillation for Efficient Speculative Decoders [arXiv] [OpenReview]

  Yuezhou Hu*, Jiaxin Guo*, Xinyu Feng, Tuo Zhao

  Neural Information Processing Systems (NeurIPS), 2025 (Spotlight)

Preprints

• Identifying Sensitive Weights via Post-quantization Integral [arXiv]

  Yuezhou Hu, Weiyu Huang, Zichen Liang, Chang Chen, Jintao Zhang, Jun Zhu, Jianfei Chen

• CAST: Continuous and Differentiable Semi-Structured Sparsity-Aware Training for Large Language Models [arXiv]

  Weiyu Huang, Yuezhou Hu, Jun Zhu, Jianfei Chen

• A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attention

  Jintao Zhang, Rundong Su, Chunyu Liu, Jia Wei, Ziteng Wang, Pengle Zhang, Haoxu Wang, Huiqiang Jiang, Haofeng Huang, Chendong Xiang, Haocheng Xi, Shuo Yang, Xingyang Li, Yuezhou Hu, Tianyu Fu, Tianchen Zhao, Yicheng Zhang, Boqun Cao, Youhe Jiang, Chang Chen, Kai Jiang, Huayu Chen, Min Zhao, Xiaoming Xu, Yi Wu, Fan Bao, Jun Zhu, Jianfei Chen

• ParallelBench: Understanding the Trade-offs of Parallel Decoding in Diffusion LLMs [arXiv]

  Wonjun Kang, Kevin Galim, Seunghyuk Oh, Minjae Lee, Yuchen Zeng, Shuibai Zhang, Coleman Hooper, Yuezhou Hu, Hyung Il Koo, Nam Ik Cho, Kangwook Lee

• Arbitrage: Efficient Reasoning via Advantage-Aware Speculation [arXiv] [Project page]

  Monishwaran Maheswaran*, Rishabh Tiwari*, Yuezhou Hu*, Kerem Dilmen, Coleman Hooper, Haocheng Xi, Nicholas Lee, Mehrdad Farajtabar, Michael Mahoney, Kurt Keutzer, Amir Gholami