From Entity-Centric to Goal-Oriented Graphs: Enhancing LLM Knowledge Retrieval in Minecraft

TL;DR

This paper introduces Goal-Oriented Graphs (GoGs), a new framework that improves LLM's procedural reasoning in Minecraft by explicitly modeling goal dependencies, outperforming existing retrieval methods.

Contribution

The paper presents a novel Goal-Oriented Graphs framework that enhances LLM reasoning by explicitly modeling goal dependencies for better multi-step planning.

Findings

GoG significantly improves procedural reasoning in Minecraft.

GoG outperforms GraphRAG and other baselines.

Explicit goal modeling aids in coherent multi-step planning.

Abstract

Large Language Models (LLMs) demonstrate impressive general capabilities but often struggle with step-by-step procedural reasoning, a critical challenge in complex interactive environments. While retrieval-augmented methods like GraphRAG attempt to bridge this gap, their fragmented entity-relation graphs hinder the construction of coherent, multi-step plans. In this paper, we propose a novel framework based on Goal-Oriented Graphs (GoGs), where each node represents a goal and edges encode logical dependencies between them. This structure enables the explicit retrieval of causal reasoning paths by identifying a high-level goal and recursively retrieving its prerequisites, forming a coherent chain to guide the LLM. Through extensive experiments on the Minecraft testbed, a domain that demands robust multi-step planning and provides rich procedural knowledge, we demonstrate that GoG substantially improves procedural reasoning and significantly outperforms GraphRAG and other state-of-the-art baselines.