Shiyuan Huang

prof_pic.jpg

I am a fifth-year Ph.D. candidate in Computer Science at DVMM Lab, Columbia University, advised by Prof. Shih-Fu Chang. My research interest is computer vision under low-resource scenarios. Specifically, I work actively on efficient multi-modal understanding, compressed video understanding and few-shot learning.

Prior to the doctoral program, I obtained my B.S. in Electrical Engineering and B.S. in Applied Mathematics at UC San Diego in 2018. I was fortunate to work with Prof. Truong Nguyen and Prof. Jorge Cortés.

News

May 23, 2022 I started my summer internship at Amazon Alexa AI on video summarization.
May 17, 2021 I started my summer internship at Amazon Alexa AI on task-specific video compression.
May 26, 2020 I started my summer internship at Adobe Research on multi-take video alignment.

Selected Publications

  1. CVPR
    Task-Adaptive Negative Envision for Few-Shot Open-Set Recognition
    Shiyuan Huang*, Jiawei Ma*, Guangxing Han, and Shih-Fu Chang
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2022
    Paper
  2. CVPR
    Few-Shot Object Detection With Fully Cross-Transformer
    Guangxing Han, Jiawei Ma, Shiyuan Huang, Long Chen, and Shih-Fu Chang
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2022
    Paper
  3. AAAI
    Meta faster r-cnn: Towards accurate few-shot object detection with attentive feature alignment
    Guangxing Han, Shiyuan Huang, Jiawei Ma, Yicheng He, and Shih-Fu Chang
    Thirty-Sixth AAAI Conference on Artificial Intelligence (AAAI) 2021
    Paper
  4. ICCV
    Query adaptive few-shot object detection with heterogeneous graph convolutional networks
    Guangxing Han, Yicheng He, Shiyuan Huang, Jiawei Ma, and Shih-Fu Chang
    In Proceedings of the IEEE/CVF International Conference on Computer Vision 2021
    Paper
  5. arxiv
    Flow-distilled IP two-stream networks for compressed video action recognition
    Shiyuan Huang, Xudong Lin, Svebor Karaman, and Shih-Fu Chang
    arXiv preprint arXiv:1912.04462 2019
    Paper