RRCM: Ranking-Driven Retrieval over Collaborative and Meta Memories for LLM Recommendation

2026-05-07

AI Summary

This work addresses the challenge in large language model (LLM)-based recommender systems where heterogeneous evidence—such as collaborative behavior and item metadata—is difficult to dynamically integrate into decision-relevant contexts and often constrained by input length limits. To overcome this, the authors propose a ranking-driven joint retrieval-and-reasoning framework that begins with lightweight user histories and employs an end-to-end optimized policy to dynamically determine whether and how to retrieve either collaborative or metadata memories. These memories are uniformly represented and accessed in natural language form. Crucially, memory retrieval is directly driven by final recommendation quality, eliminating the need for handcrafted rules or fixed pipelines and enabling adaptive, interleaved utilization of both memory types. Experiments demonstrate that the proposed approach significantly outperforms existing baselines and LLM-based recommendation methods in both recommendation accuracy and contextual efficiency.

Abstract

Large Language Models (LLMs) have emerged as a promising paradigm for next-generation recommender systems, offering strong semantic understanding and natural-language reasoning abilities. Despite recent progress, current LLM-based recommenders still face key challenges in constructing decision-relevant contexts from heterogeneous evidence. First, existing methods often rely on fixed context construction strategies: collaborative behavioral evidence and item-side metadata are typically incorporated through predefined prompts, static retrieval pipelines, or handcrafted injection mechanisms, making it difficult to determine what information is truly beneficial for each instance. Second, heterogeneous evidence introduces a severe context-efficiency bottleneck. Rich metadata and collaborative interaction records can quickly overwhelm the context window, while aggressive compression or heuristic filtering may discard fine-grained evidence critical for accurate recommendation. To address these challenges, we propose RRCM, a ranking-driven retrieval-and-reasoning framework over collaborative and metadata memories for LLM-based agentic recommendation. RRCM starts from a lightweight user-history context and learns whether to recommend directly, retrieve collaborative evidence, retrieve item metadata, or interleave both through reasoning. Both memories are represented in natural language and accessed through a unified retrieval interface, enabling flexible evidence acquisition without handcrafted CF injection or fixed retrieval rules. We optimize this memory-reading policy with an outcome-only ranking reward, instantiated using group relative policy optimization, so that retrieval decisions are directly driven by final top-k recommendation quality. Extensive experiments show that RRCM significantly outperforms traditional baselines and diverse LLM-based recommendation approaches.

Problem

• LLM-based recommendation
• heterogeneous evidence
• context construction
• collaborative memory
• metadata retrieval

Innovation

• ranking-driven retrieval
• collaborative memory
• meta memory
• LLM-based recommendation
• context-efficient reasoning