Event
FISRF: Adaptively Guiding Robotic Task Planning Reflection with a Mentor Architecture
1 reports · 1 sources
Qingmiao Liang and other authors propose FISRF (Flexible Instructional Scaffolding Reflection Framework), which adopts an agent–mentor architecture to estimate the robot's zone of proximal development (ZPD) in task planning and, on that basis, adaptively selects reflection strategies and mentor guidance, in order to mitigate the accumulation of erroneous experience during early interactions. The authors state that experiments on ALFWorld and a real robot platform show that this method reduces the average number of error-correction rounds by 37% and redundant reasoning by 25.1%, and increases overall task success rate by 8%. These figures come from the paper's report and await independent verification.
Report timeline
- FISRF: An Adaptive Reflection-Guided Framework for LLM-Based Robot Task Planning
FISRF proposes an agent-and-mentor architecture that estimates a robot's Zone of Proximal Development (ZPD) in task planning and adaptively selects reflection strategies and mentor guidance, mitigating the accumulation of erroneous experience during early interactions. Experiments in ALFWorld and on real-world robotic platforms show that the method reduces average error-correction rounds by 37% and redundant reasoning by 25.1%, while improving overall task success rate by 8%.
arXiv cs.RO ·