Chenglin Yang

Chenglin Yang

Research Scientist

ByteDance


I am currently a Research Scientist at ByteDance.

I received my Ph.D. in Computer Sciense, advised by Bloomberg Distinguished Professor Alan Yuille, and M.S. in Robotics from Johns Hopkins University.

Before that, I obtained my B.E. in Engineering Mechanics from Beijing Jiaotong University, where I was honored to work with Changjiang Distinguished Professor Yue-Sheng Wang.

During my graduate studies, I have interned at Google DeepMind, Google Research, and Adobe.

My Chinese name is 杨程麟.

Interests
  • Computer Vision
  • Vision Language Model
  • Neural Network Architecture
  • Reinforcement Learning
  • Knowledge Distillation
Education
  • Ph.D. in Computer Science

    Johns Hopkins University, 2019 - 2024

  • M.S. in Robotics

    Johns Hopkins University, 2017 - 2019

  • B.E. in Engineering Mechanics

    Beijing Jiaotong University, 2013 - 2017

Publications

(2023). IG Captioner: Information Gain Captioners are Strong Zero-shot Classifiers. In ECCV24.

PDF

(2022). MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models. In ICLR23.

PDF Code

(2022). Lite Vision Transformer with Enhanced Self-Attention. In CVPR22.

PDF Code

(2020). PatchAttack: A Black-box Texture-based Attack with Reinforcement Learning. In ECCV20.

PDF Code

(2019). Snapshot Distillation: Teacher-Student Optimization in One Generation. In CVPR19.

PDF

(2019). Training Deep Neural Networks in Generations: A More Tolerant Teacher Educates Better Students. In AAAI19.

PDF