RoboTTT Enables Long-Memory Robot Learning
DrJimFan · x · 2026-07-15
Introducing RoboTTT: applying test-time training to robotic policies. By performing gradient updates on a small kernel network during inference, it compresses historical experience into the model weights.
Key results include:
- Natively extends the robot's context length to 8,000 timesteps, equivalent to about 5 minutes of "muscle memory," while keeping inference costs constant.
- Unlike older policies that process a few frames at a time and quickly forget, this approach allows the robot to learn continuously and improve even after deployment.
- Supports one-shot in-context learning: after a single human video demonstration, the robot can imitate previously unseen circuit board assembly configurations.
- If an error occurs during execution, the robot can incorporate the "corrective action" into its context, enabling self-correction.
The authors also report a new context scaling curve: from 128 to 8K timesteps, closed-loop performance consistently improves without saturating; 8K context pre-training boosts performance by 62% compared to 1K.
Related event: RoboTTT brings test-time training to robot policies(6 posts)→
More from Embodied
- LIBERO-MAX: Robot Policies Drop 11–26 Points When the World Changes Mid-Task — drmapavone · 2026-10-27
- Photos from ROSCon UK: BBC's robotic whale, da Vinci surgical robot, PR2 and Baxter — kamathsblog · 2026-10-27
- Agility's Digit robot to expand into two more GXO US warehouses from early 2027 — jonstephens85 · 2026-10-27
- Real2Sem2Real turns one phone video into 2,000+ robot demos, matching 150 teleop demos — Scobleizer · 2026-10-27
- Wayve CEO argues end-to-end deep learning should power driving, safety and world models — alexgkendall · 2026-10-27
- Mundo turns a few room photos into an explorable, simulatable 3D world — at 1.5 FPS — mark_k · 2026-10-27