Zeyu Wang

I am a PhD student at Princeton University, where I am a member of Princeton Visual AI Lab. I am advised by Professor Olga Russakovsky, and I also work closely with Professor Karthik Narasimhan. I did my bachelors at Fudan University.

I work at the intersection of computer vision and natural language processing, with the goal to build intelligent agents that can both see and speak. To this end, I am currently working on grounding language concepts in vision domain. Besides vision and language, I also have interests in self-supervised learning and continual learning.

Email  /  GitHub

profile photo
Research
spice-u Towards Unique and Informative Captioning of Images
Zeyu Wang, Berthy Feng, Karthik Narasimhan and Olga Russakovsky
ECCV, 2020
paper / code / video

We propose SPICE-U, a new metric for captioning evaluation which explicitly captures uniqueness of captions, and an inference time re-ranking technique for generating more descriptive captions.

spice-u Towards Fairness In Visual Recognition: Effective Strategies For Bias Mitigation
Zeyu Wang, Klint Qinami, Ioannis C. Karakozis, Kyle Genova, Prem Nair, Kenji Hata and Olga Russakovsky
CVPR, 2020
paper / code / video

We aim to increase the fairness of visual recognition models by providing a systematic study of existing and alternative bias mitigation techniques.

Teaching
  • Teaching Assistant, ELE 364 Machine Learning for Predictive Data Analytics
  • Teaching Assistant, ELE 535 Machine Learning and Pattern Recognition

Template from here.