Yuancheng Xu

Research Scientist
Netflix Eyeline Studio
Email: xuyuancheng0 [at] gmail [dot] com
About me

My name is Yuancheng Xu (徐源诚) and I am a Research Scientist at Netflix Eyeline Lab, where I develop foundational AI models with advanced multimodal understanding, reasoning, interaction, and generation capabilities. I earned my Ph.D. in 2025 from the University of Maryland, College Park, under the supervision of Furong Huang.

Research



[GenARM: Reward Guided Generation with Autoregressive Reward Model for Test-Time Alignment]
GenARM: Reward Guided Generation with Autoregressive Reward Model for Test-Time Alignment
Yuancheng Xu, Udari Madhushani Sehwag, Alec Koppel, Sicheng Zhu, Bang An, Furong Huang, Sumitra Ganesh
International Conference on Learning Representations (ICLR), 2025
  • Propose the Autoregressive Reward Model to efficiently guide next-token generation.
[Paper] [Code] [Project Page]










[Exploring and Exploiting Decision Boundary Dynamics for Adversarial Robustness]
Exploring and Exploiting Decision Boundary Dynamics for Adversarial Robustness
Yuancheng Xu, Yanchao Sun, Micah Goldblum, Tom Goldstein, Furong Huang
International Conference on Learning Representations (ICLR), 2023
  • Mathematical characterization of how decison boundary moves.
  • Analytic expression for the margin's gradient. Directly applicable to robust training.
[Paper] [Code]