REFLECT: Summarizing Robot Experiences for Failure Explanation and Correction
This work addresses the need for explainable and robust robotic systems by enabling automated failure analysis and correction, representing an incremental improvement through the novel integration of LLMs with robot experiences.
The authors tackled the problem of automatically detecting and analyzing robot failures by introducing REFLECT, a framework that uses hierarchical summaries of past experiences to query Large Language Models for failure reasoning, which then guides correction planning; they demonstrated its ability to generate informative explanations that assist in successful corrections, as evaluated on the RoboFail dataset.
The ability to detect and analyze failed executions automatically is crucial for an explainable and robust robotic system. Recently, Large Language Models (LLMs) have demonstrated strong reasoning abilities on textual inputs. To leverage the power of LLMs for robot failure explanation, we introduce REFLECT, a framework which queries LLM for failure reasoning based on a hierarchical summary of robot past experiences generated from multisensory observations. The failure explanation can further guide a language-based planner to correct the failure and complete the task. To systematically evaluate the framework, we create the RoboFail dataset with a variety of tasks and failure scenarios. We demonstrate that the LLM-based framework is able to generate informative failure explanations that assist successful correction planning.