Annie Xie
I am a research scientist at Google DeepMind. My current focus is on building agents for the physical world through embodied and multimodal reasoning.
I received my Ph.D. in CS in 2024 from Stanford, where I was advised by Chelsea Finn. My thesis was on reinforcement learning algorithms for efficient exploration and adaptation in dynamic environments. I also completed internships at Google Brain and Meta's FAIR team working on reinforcement learning for robotic control.
Email  / 
CV  / 
Google Scholar
|
|
Select Papers
|
Gemini Robotics: Bringing AI into the Physical World
Gemini Robotics Team
[arxiv] [blog post]
|
Learning to Explore in POMDPs with Informational Rewards
Annie Xie*,
Logan M Bhamidipaty*,
Evan Z Liu,
Joey Hong,
Sergey Levine,
Chelsea Finn
International Conference on Machine Learning (ICML), 2024
[arxiv]
|
Decomposing the Generalization Gap in Imitation Learning for Visual Robotic Manipulation
Annie Xie*,
Lisa Lee*,
Ted Xiao,
Chelsea Finn
International Conference on Robotics and Automation (ICRA), 2024
[arxiv][webpage]
|
Supervised Pretraining Can Learn In-Context Reinforcement Learning
Jonathan N. Lee*,
Annie Xie*,
Aldo Pacchiano,
Yash Chandak,
Chelsea Finn,
Ofir Nachum,
Emma Brunskill
Conference on Neural Information Processing Systems (NeurIPS), 2023 (Spotlight)
[arxiv]
|
|