BAR: A Backward Reasoning based Agent for Complex Minecraft Tasks

2025-05-20

AI Summary

Forward reasoning paradigms often fail in complex tasks due to a perceptual gap between the initial and goal states, leading to brittle planning. Method: This paper proposes BAR, a backward-reasoning agent designed for the Minecraft virtual environment. BAR introduces a terminal-state-driven recursive goal decomposition mechanism, integrated with state-consistency modeling and a phased memory architecture, enabling efficient backward derivation of executable action sequences from the goal. Contribution/Results: BAR overcomes key limitations of conventional forward planning—namely, sensitivity to initial conditions and weak modeling of long-horizon dependencies. Experiments demonstrate that BAR significantly improves planning robustness, state consistency, and reasoning efficiency on complex Minecraft tasks, achieving an average success rate 32.7% higher than the best forward-baseline method.

Technology Category

• Multiagent Systems: Multiagent Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Large language model (LLM) based agents have shown great potential in following human instructions and automatically completing various tasks. To complete a task, the agent needs to decompose it into easily executed steps by planning. Existing studies mainly conduct the planning by inferring what steps should be executed next starting from the agent's initial state. However, this forward reasoning paradigm doesn't work well for complex tasks. We propose to study this issue in Minecraft, a virtual environment that simulates complex tasks based on real-world scenarios. We believe that the failure of forward reasoning is caused by the big perception gap between the agent's initial state and task goal. To this end, we leverage backward reasoning and make the planning starting from the terminal state, which can directly achieve the task goal in one step. Specifically, we design a BAckward Reasoning based agent (BAR). It is equipped with a recursive goal decomposition module, a state consistency maintaining module and a stage memory module to make robust, consistent, and efficient planning starting from the terminal state. Experimental results demonstrate the superiority of BAR over existing methods and the effectiveness of proposed modules.

Problem

• Forward reasoning fails for complex Minecraft tasks
• Big perception gap between initial state and goal
• Backward reasoning improves planning for task completion

Innovation

• Backward reasoning from terminal state
• Recursive goal decomposition module
• State consistency maintaining module