publications

(*) denotes equal contribution

2024

  1. arXiv
    Learning to (Learn at Test Time): RNNs with Expressive Hidden States
    Yu Sun*, Xinhao Li*, Karan Dalal*, Jiarui Xu, Arjun Vikram, Genghan Zhang, Yann Dubois, Xinlei Chen, Xiaolong Wang, Sanmi Koyejo, Tatsunori Hashimoto, and Carlos Guestrin
    2024
    PDF
  2. arXiv
    MoA: Mixture of Sparse Attention for Automatic Large Language Model Compression
    Tianyu Fu*, Haofeng Huang*, Xuefei Ning*, Genghan Zhang, Boju Chen, Tianqi Wu, Hongyi Wang, Zixiao Huang, Shiyao Li, Shengen Yan, Guohao Dai, Huazhong Yang, and Yu Wang
    2024
    PDF
  3. COLM
    CATS: Contextually-Aware Thresholding for Sparsity in Large Language Models
    Je-Yong Lee*, Donghyun Lee*, Genghan Zhang, Mo Tiwari, and Azalia Mirhoseini
    2024
    PDF
  4. PLDI
    Compilation of Modular and General Sparse Workspaces
    Genghan Zhang, Olivia Hsu, and Fredrik Kjolstad
    Proceedings of the ACM on Programming Languages 2024
    PDF
  5. arXiv
    GeoT: Tensor Centric Library for Graph Neural Network via Efficient Segment Reduction on GPU
    Zhongming Yu, Genghan Zhang, Hanxian Huang, Xin Chen, and Jishen Zhao
    arXiv preprint 2024
    PDF

2023

  1. CCFTHPC
    Sgap: towards efficient sparse tensor algebra compilation for GPU
    Genghan Zhang, Yuetong Zhao, Yanting Tao, Zhongming Yu, Guohao Dai, Sitao Huang, Yuan Wen, Pavlos Petoumenos, and Yu Wang
    CCF Transactions on High Performance Computing 2023
    PDF
  2. MLSys
    HyperGef: A Framework Enabling Efficient Fusion for Hypergraph Neural Network on GPUs
    Zhongming Yu, Guohao Dai, Shang Yang, Genghan Zhang, Hengrui Zhang, Feiwen Zhu, June Yang, Jishen Zhao, and Yu Wang
    Proceedings of Machine Learning and Systems 2023
    PDF