MoM: Mixtures of Scenario-Aware Document Memories for Retrieval-Augmented Generation Systems

Jihao Zhao, Zhiyuan Ji, Simin Niu, Hanyu Wang, Feiyu Xiong, Zhiyu li

Keywords: MoM, Document memory extraction, Text chunking, Proactive understanding, Large language models

Abstract: The traditional retrieval-augmented generation (RAG) paradigm, which typically engages in the comprehension of relevant text chunks in response to received queries, inherently restricts both the depth of knowledge internalization and reasoning capabilities. To address this limitation, our research transforms the text processing in RAG from passive chunking to proactive understanding, defining this process as document memory extraction with the objective of simulating human cognitive processes during reading. Building upon this, we propose the Mixtures of scenario-aware document Memories (MoM) framework, engineered to efficiently handle documents from multiple domains and train small language models (SLMs) to acquire the ability to proactively explore and construct document memories. The MoM initially instructs large language models (LLMs) to simulate domain experts in generating document logical outlines, thereby directing structured chunking and core content extraction. It employs a multi-path sampling and multi-perspective evaluation mechanism, specifically designing comprehensive metrics that represent chunk clarity and extraction completeness to select the optimal document memories. Additionally, to infuse deeper human-like reading abilities during the training of SLMs, we incorporate a reverse reasoning strategy, which deduces refined expert thinking paths from high-quality outcomes. Finally, leveraging diverse forms of content generated by MoM, we develop a three-layer document memory retrieval mechanism, which is grounded in our theoretical proof from the perspective of probabilistic modeling: compared to the strategy of fusing embeddings prior to retrieval, independently retrieving memories from each layer and subsequently fusing them can more effectively reduce information loss. Extensive experimental results across three distinct domains demonstrate that the MoM framework not only resolves text chunking challenges in existing RAG systems, providing LLMs with semantically complete document memories, but also paves the way for SLMs to achieve human-centric intelligent text processing.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 24467