Echo teaches VLA robots to detect task success and failure without frame-level labels
Yufeng Liu · hf · 2026-10-26
- Echo, from SJTU, Spirit AI, Tsinghua and others, enables Vision-Language-Action models to recognize task completion and failure so robots can stop on success and recover or replan after failure.
- Stop and failure signals are added as two extra dimensions of each action chunk and denoised together with robot actions, within a single compact VLA—no frame-level progress annotations, external monitors, or architectural changes.
- Signals are learned from complementary successful, failed, and recovery trajectories, then guide action generation.
More from Embodied
- HKU/ZJU/Fudan with Lightspeed launch HiSTEVE: hierarchical VLA agent playing Minecraft via native mouse and keyboard — tencent · 2026-10-26
- Hobbyist teaches StackChan robot to read scripts with Grok and Hermes-Agent — alexcovo_eth · 2026-10-26
- 10 wild humanoid robots: ex-Boston Dynamics droid butler, car-lifting bodyguard bot, ladder-climbing G1 — minchoi · 2026-10-26
- KAIST's humanoid v0.9 pushes 35 kg and lifts 20 kg, then moonwalks — minchoi · 2026-10-26
- Alibaba DAMO releases 16B unified embodied model RynnUnify, 51.5% on RoboCasa365 — Minghao Zhu · 2026-10-26
- HoloBoxing teaches humanoid boxing with just 7 minutes of single-GPU training — Xihan Ma · 2026-10-26