MemORAI: Memory Organization and Retrieval via Adaptive Graph Intelligence for LLM Conversational Agents

2026-05-02

AI Summary

This work addresses the lack of effective long-term memory mechanisms in large language models for personalized dialogue, where existing graph-based memory systems suffer from information dilution, missing provenance, and rigid retrieval. The authors propose a novel memory architecture that employs two-stage compression for selective memory filtering, constructs a multi-relational knowledge graph enabling turn-level fact tracing, and introduces a query-adaptive subgraph retrieval mechanism based on dynamically weighted PageRank. Evaluated on the LOCOMO and LongMemEval benchmarks, the approach significantly improves memory retrieval accuracy and the quality of personalized responses, achieving state-of-the-art performance.

Abstract

Large Language Models (LLMs) lack persistent memory for long-term personalized conversations. Existing graph-based memory systems suffer from information dilution, absent provenance tracking, and uniform retrieval that ignores query context. We introduce MemORAI (Memory Organization and Retrieval via Adaptive Graph Intelligence), a framework that integrates three innovations: selective memory filtering with dual-layer compression to retain user-persona-relevant content, a provenance-enriched multi-relational graph tracking factual origins at the turn level, and query-adaptive subgraph retrieval with Dynamic Weighted PageRank that applies query-conditioned edge weighting. Evaluated on LOCOMO and LongMemEval benchmarks, MemORAI achieves state-of-the-art performance in memory retrieval and personalized response generation, demonstrating that selective storage, enriched representation, and adaptive retrieval are essential for coherent, personalized LLM agents.

Problem

• persistent memory
• personalized conversation
• information dilution
• provenance tracking
• adaptive retrieval

Innovation

• selective memory filtering
• provenance-enriched graph
• query-adaptive retrieval
• Dynamic Weighted PageRank
• multi-relational memory graph