Sergey Levine
UC Berkeley professor and co-author on QAM, IQL, SERL, RLDG, AWAC — foundational algorithmic work underlying the LWD system.
“Chelsea Finn & Sergey Levine — authors of the π0/π0.5 VLA baselines and much of the underlying data ecosystem”
Source→“Chelsea Finn (π0, ALOHA, Mobile ALOHA), Shuran Song (ACT, UMI, Diffusion Policy), Sergey Levine, Pieter Abbeel (GELLO, MuJoCo Playground), and Russ Tedrake (UMI)”
Source→“Sergey Levine... Referenced as co-author on π0.5, π0.7, RT-2, SayCan, and ECoT”
Source→“Sergey Levine... One of the most influential researchers in robot learning and reinforcement learning. His involvement signals that this approach is grounded in rigorous RL theory and state-of-the-art robotic control”
Source→“Authors: Andy Tang, William Chen, Andrew Wagenmaker, Chelsea Finn, Sergey Levine (Stanford + UC Berkeley). Date: June 2025. arXiv: 2606.13675.”
Source→“FRS is an inference-time mechanism to unlock that latent knowledge without retraining the base model.”
Source→“We use OpenPi's π0.5-LIBERO... For all others, we use π0.5 fine-tuned by Jain et al.”
Source→“Levine's group is systematically building the infrastructure for generalist robot policies that can be rapidly adapted — FRS is one piece of that stack.”
Source→“Sergey Levine (SAC, π₀)... These are the intellectual ancestors of the approach.”
Source→“Multiple citations throughout; co-author on QAM [31], IQL [22], SERL [40], RLDG [56]”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.