GroupRAG: Cognitively Inspired Group-Aware Retrieval and Reasoning via Knowledge-Driven Problem Structuring

TL;DR

GroupRAG introduces a cognitively inspired framework that enhances language model reasoning by identifying problem structure through knowledge-driven grouping, leading to improved performance on complex question-answering tasks.

Contribution

It proposes a novel group-aware retrieval and reasoning method based on problem structure, inspired by cognitive science, to improve knowledge integration and reasoning in language models.

Findings

GroupRAG outperforms RAG and CoT baselines on MedQA.

Explicit modeling of problem structure improves reasoning robustness.

Knowledge-driven grouping enables multi-start reasoning processes.

Abstract

The performance of language models is commonly limited by insufficient knowledge and constrained reasoning. Prior approaches such as Retrieval-Augmented Generation (RAG) and Chain-of-Thought (CoT) address these issues by incorporating external knowledge or enforcing linear reasoning chains, but often degrade in real-world settings. Inspired by cognitive science, which characterizes human problem solving as search over structured problem spaces rather than single inference chains, we argue that inadequate awareness of problem structure is a key overlooked limitation. We propose GroupRAG, a cognitively inspired, group-aware retrieval and reasoning framework based on knowledge-driven keypoint grouping. GroupRAG identifies latent structural groups within a problem and performs retrieval and reasoning from multiple conceptual starting points, enabling fine-grained interaction between the two processes. Experiments on MedQA show that GroupRAG outperforms representative RAG- and CoT-based baselines. These results suggest that explicitly modeling problem structure, as inspired by human cognition, is a promising direction for robust retrieval-augmented reasoning.