Technology Category
Multiagent Systems: Multiagent PlanningPlanning, Routing, and Scheduling: Planning with Language ModelsHumans and AI: Human-Aware Planning and Behavior Prediction
Don't Just Follow MLLM Plans: Robust and Efficient Planning for Open-world Agents
🤖 AI Summary
In open-world settings, the unreliability of LLMs’ prior knowledge and environmental interaction uncertainty lead to fragile multi-step task planning and inefficient exploration. Method: We propose the first autonomous agent framework that learns planning knowledge from scratch, integrating adaptive dependency modeling, fine-grained failure-aware operational memory, and difficulty-driven active exploration—combined with reinforcement learning, memory-augmented architectures, and LLM-guided reasoning. Contribution/Results: Without relying on external knowledge or idealized assumptions, our framework significantly improves planning robustness and sample efficiency. On two open-world benchmarks, it achieves substantially higher planning success rates than prior methods and, for the first time, solves high-difficulty late-stage item acquisition tasks—demonstrating the feasibility of end-to-end learning of reliable planning knowledge directly from real-world interactions.
Technology Category
Application Category
📝 Abstract
Developing autonomous agents capable of mastering complex, multi-step tasks in unpredictable, interactive environments presents a significant challenge. While Large Language Models (LLMs) offer promise for planning, existing approaches often rely on problematic internal knowledge or make unrealistic environmental assumptions. Although recent work explores learning planning knowledge, they still retain limitations due to partial reliance on external knowledge or impractical setups. Indeed, prior research has largely overlooked developing agents capable of acquiring planning knowledge from scratch, directly in realistic settings. While realizing this capability is necessary, it presents significant challenges, primarily achieving robustness given the substantial risk of incorporating LLMs' inaccurate knowledge. Moreover, efficiency is crucial for practicality as learning can demand prohibitive exploration. In response, we introduce Robust and Efficient Planning for Open-world Agents (REPOA), a novel framework designed to tackle these issues. REPOA features three key components: adaptive dependency learning and fine-grained failure-aware operation memory to enhance robustness to knowledge inaccuracies, and difficulty-based exploration to improve learning efficiency. Our evaluation in two established open-world testbeds demonstrates REPOA's robust and efficient planning, showcasing its capability to successfully obtain challenging late-game items that were beyond the reach of prior approaches.
Problem
Research questions and friction points this paper is trying to address.
Developing autonomous agents for unpredictable environments
Reducing reliance on inaccurate LLM knowledge
Improving planning efficiency in open-world tasks
Innovation
Methods, ideas, or system contributions that make the work stand out.
Adaptive dependency learning for robustness
Fine-grained failure-aware operation memory
Difficulty-based exploration for efficiency
🔎 Similar Papers
No similar papers found.
