Research: Huizi's research focuses on software and hardware co-design for efficient deep learning. He is interested in algorithms to compress and then accelerate deep neural networks on general purpose processors or custom hardware. Recently he is interested in deep learning for video, which involves algorithm design to exploit temporal locality in video and efficiently execute the model on hardware.
Bio: Huizi is a PhD candidate in Stanford's Electrical Engineering department. He received B.E. in Electronic Engineering and B.S. in Mathematics from Tsinghua University.