Xingyi Yang

I am an incoming Ph.d student at National University of Singapore(NUS) at ECE department. I am now working under the supervision of Prof.Xinchao Wang. I got my master’s degree in Electrical and Computer Engineering from University of California, San Diego in 2021 and my bachelor degree in Computer Science and Engineering from Southeast University, China in 2019. I am also fortunate to work with Prof.Pengtao Xie, Prof.Rose Yu, Prof.Fenglong ma and Prof.Yining Hu. I used to be a visiting student at University of Ottawa, supervised by Prof.Robert Laganière.

My current research interest can be summarized into 3M scheme:

  • Statistical Machine learning. Focus on generative modeling, trust-worthy learning (interpretability and robustness) and graph learning.
  • Minimal supervision. Focus on self-supervised & semi-supervised & weak-supervised learning, transfer learning, knowledge-grounded learning.
  • Medical data analysis, especially during the outbreak of the pandemic.


I believe in Slow Science


  • [2021/05] Two paper accepted by ACL2021, one about Medical report generation and one on COVID-19 Dialogues.
  • [2021/02] One paper accepted by CVPR2021, DSRNA: Differentiable Search of Robust Neural Architectures.
  • [2021/01] One paper accepted by ICASSP2021, Kalman Optimizer for Consistent Gradient Descent.
  • [2020/09] I am co-organizing NeurIPS 2020 workshop “Self-Supervised Learning – Theory and Practice”, serving as the Workflow Chair.
  • [2020/09] One oral paper accepted by ISVC2020.
  • [2020/07] One paper accepted by ECCV2020, Single Viaew Metrology in the Wild