13 papers
mjlab: A Lightweight Framework for GPU-Accelerated Robot Learning
Kevin Zakka, Qiayuan Liao, Brent Yi +3
We present mjlab, a lightweight, open-source framework for robot learning that combines GPU-accelerated simulation with composable environments and minimal setup friction. mjlab ad…
Flow Policy Gradients for Robot Control
Brent Yi, Hongsuk Choi, Himanshu Gaurav Singh +9
Likelihood-based policy gradient methods are the dominant approach for training robot control policies from rewards. These methods rely on differentiable action likelihoods, which…
Coordinated Humanoid Manipulation with Choice Policies
Haozhi Qi, Yen-Jen Wang, Toru Lin +4
Humanoid robots hold great promise for operating in human-centric environments, yet achieving robust whole-body coordination across the head, hands, and legs remains a major challe…
Cameras as Relative Positional Encoding
Ruilong Li, Brent Yi, Junchen Liu +3
Transformers are increasingly prevalent for multi-view computer vision tasks, where geometric relationships between viewpoints are critical for 3D perception. To leverage these rel…
Eye, Robot: Learning to Look to Act with a BC-RL Perception-Action Loop
Justin Kerr, Kush Hari, Ethan Weber +5
Humans do not passively observe the visual world -- we actively look in order to act. Motivated by this principle, we introduce EyeRobot, a robotic system with gaze behavior that e…
Flow Matching Policy Gradients
David McAllister, Songwei Ge, Brent Yi +5
Flow-based generative models, including diffusion models, excel at modeling continuous distributions in high-dimensional spaces. In this work, we introduce Flow Policy Optimization…