Understand Then Memory: A Cognitive Gist-Driven RAG Framework with Global Semantic Diffusion

TL;DR

Inspired by the human episodic memory mechanism, CogitoRAG is proposed, a RAG framework that simulates human cognitive memory processes that significantly outperforms state-of-the-art RAG methods, showcasing superior capabilities in complex knowledge integration and reasoning.

Abstract

Retrieval-Augmented Generation (RAG) effectively mitigates hallucinations in LLMs by incorporating external knowledge. However, the inherent discrete representation of text in existing frameworks often results in a loss of semantic integrity, leading to retrieval deviations. Inspired by the human episodic memory mechanism, we propose CogitoRAG, a RAG framework that simulates human cognitive memory processes. The core of this framework lies in the extraction and evolution of the Semantic Gist. During the offline indexing stage, CogitoRAG first deduces unstructured corpora into gist memory corpora, which are then transformed into a multi-dimensional knowledge graph integrating entities, relational facts, and memory nodes. In the online retrieval stage, the framework handles complex queries via Query Decomposition Module that breaks them into comprehensive sub-queries, mimicking the cognitive decomposition humans employ for complex information. Subsequently, Entity Diffusion Module performs associative retrieval across the graph, guided by structural relevance and an entity-frequency reward mechanism. Furthermore, we propose the CogniRank algorithm, which precisely reranks candidate passages by fusing diffusion-derived scores with semantic similarity. The final evidence is delivered to the generator in a passage-memory pairing format, providing high-density information support. Experimental results across five mainstream QA benchmarks and multi-task generation on GraphBench demonstrate that CogitoRAG significantly outperforms state-of-the-art RAG methods, showcasing superior capabilities in complex knowledge integration and reasoning.

Figures (2)

• Figure 1: (a) Vector-indexed RAG: relies on local semantic matching of document chunks. (b) Knowledge Graph-indexed RAG: captures explicit entity links, limited to local triple reasoning. (c) Our CogitoRAG: extracts gist memory to capture explicit facts and implicit logic, enabling holistic understanding.
• Figure 2: The framework follows the "Understand Then Memory" paradigm: (1) Offline Indexing, it distills explicit facts and implicit relational logic from raw passages to construct a multi-dimensional knowledge graph that integrates semantic gist memories and contextual passages. (2) Online Retrieval, for a given query, it decomposes complex reasoning into sub-queries, then performs a unified cognitive diffusion process within the multi-dimensional knowledge graph. This enables global context-aware reranking of memories and passages via the CogniRank algorithm, delivering high-density evidential support to the generator for accurate and interpretable reasoning.