Think, Reflect, Create: Metacognitive Learning for Zero-Shot Robotic Planning with LLMs
Wenjie Lin, Jin Wei-Kocsis · 2025
Read Think, Reflect, Create: Metacognitive Learning for Zero-Shot Robotic Planning with LLMs by Wenjie Lin, Jin Wei-Kocsis (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-14899.txt.
Description
Language models can plan robot tasks from a prompt, but they stumble on new problems with no examples to copy. Wenjie Lin and co-authors built REFLEX in 2025, giving model-driven robot teams a metacognitive loop: think, reflect, then create.
The system breaks past tasks into reusable skills, reflects on why attempts fail in fresh scenarios, and assembles new solutions from the parts. On a standard benchmark and a harder new one, REFLEX beat existing baselines clearly. Strikingly, some of its winning solutions differed from the reference answers yet still worked. The authors read this as early evidence that reflection can breed genuine creativity in robot planning.
Catalog record
| CREATOR | Wenjie Lin, Jin Wei-Kocsis |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 16 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-14899