GraphThink: Graph-Enhanced LLM Thinking for Long-Horizon Embodied Task Planning
This work is significant for embodied AI researchers and developers, as it improves the robustness, generalization, and environmental awareness of LLM-based embodied agents, particularly for long-horizon tasks.
This paper addresses challenges in LLM-based embodied task planning, such as physical hallucinations and poor generalization, by introducing GraphThink. This framework uses a task graph for structured knowledge and a scene graph for environmental memory, achieving state-of-the-art performance on the ALFRED benchmark and outperforming leading API-based LLMs on long-horizon tasks.
Embodied agents using LLM-based planners often struggle with physical hallucinations, poor generalization to long-horizon tasks, and lack of environmental awareness. We propose GraphThink, a novel framework that integrates a task graph to provide structured knowledge for robust planning and a scene graph to maintain environmental memory for event-driven replanning. Specifically, the task graph guides LLM thinking through contextual prompting and iterative refinement, effectively mitigating planning hallucinations. Furthermore, within the GRPO framework, the task graph offers delicate reward design to train the LLM planner, enhancing long-horizon planning capabilities and improving generalization. Finally, an event-driven replanning module, powered by the scene graph, enables closed-loop environment awareness and error correction. GraphThink achieves state-of-the-art performance on the ALFRED benchmark. In particular, our high-level planner surpasses leading API-based LLMs on both the validation set and held-out long-horizon tasks, underscoring its robust zero-shot and few-shot capabilities. Additional evaluations further demonstrate strong out-of-distribution generalization to novel tasks and environments.