Tsinghua and AgiBot's Collective Experience Learning lets robot fleets teach each other from failures
KyeGomezB · x · 2026-10-20
Tsinghua University, AgiBot and Infinigence AI introduce Collective Experience Learning (CEL), where deployed robots share execution data to improve a common policy instead of relying on more human supervision.
- Unlike SFT (human demonstrations) or HG-DAgger (human corrections/takeovers), the experience source is the robot fleet itself, requiring minimal expert action supervision
- On failure, a robot's failure context is used to retrieve successful segments from peers in a collective experience buffer
- No extra demos or corrections needed, letting robots improve as deployment scales
Related event: Tsinghua and Partners Propose Collective Learning for Robot Fleets(2 posts)→
More from Embodied
- Microsoft launches $2,599 Surface Laptop Ultra with RTX Spark chip for local AI — tristanbob · 2026-10-23
- Benchmarking Argon and Astra on real hardware plus 1,000 sim runs: has robot manipulation been solved? — paulfinneyx · 2026-10-23
- Reversible computing veteran poses open problem: universal BARCS primitives from unbiased Josephson junctions — MikePFrank · 2026-10-23
- Insider pushes back on rumor that Figure's Hugo robot is built on a Unitree G1 — carlosdponx · 2026-10-23
- Sergey Levine: robotics is repeating LLMs' shift from fine-tuning to plain-language prompting — adityaag · 2026-10-23
- Embodied First Exam benchmarks coding agents on 1,090 tasks across 44 robots — ManlingLi_ · 2026-10-23