MemLoc improves long-term chat memory retrieval and answer quality
Where to Look and What to Use: Retrieve-Localize-Generate for Long-Term Conversational Memory Question Answering
Computation and LanguageInformation Retrieval
Summary
Remembering details from long conversations is hard because information can be scattered and mixed up. The authors created MemLoc, which breaks conversations into smaller parts, finds the most helpful bits, and guides language models to use them correctly. This method helps computers give better answers by focusing on the right pieces of past conversations. Tests showed MemLoc works better and faster than other methods.
What this means in practice
- •For chatbot developers: Improve long-term conversation memory retrieval to provide more accurate and context-aware responses in dialogue systems.
- •For customer support teams: Enhance automated support agents’ ability to reference past interactions for better problem resolution over extended conversations.
Authors
Yifan Wang, Xinkui Lin, Yongxiu Xu, Shen Gao, Ruochen Yang, Kun Huang, Yubin Wang, Jie Wu, Wei Liu, Jian Luan, Hongbo Xu, Shuo Shang
Abstract
Retrieval-augmented generation (RAG) enables large language models (LLMs) to answer questions by accessing external knowledge and has been widely adopted for long-term conversational memory question answering. However, existing methods suffer from two key challenges: (1) fragmented evidence scattered across temporally distant sessions, and (2) noisy content within retrieved sessions that triggers the lost-in-the-middle effect. To address these challenges, we propose MemLoc, a unified Retrieve-Localize-Generate framework for long-term conversational memory QA. For retrieval, MemLoc decomposes each session into multi-granularity memory units and performs query routing via an inner-memory graph with entropy-based granularity selection. It further models cross-session semantic and temporal dependencies through a cross-memory graph, enabling coarse-to-fine retrieval of top-K relevant memory candidates. For localization, we introduce a reasoning-based evidence locator trained with Self-reflective Hint Policy Optimization (SHPO), which performs progressive refinement by extracting query-relevant fragments within memory units to suppress noise and reranking across candidates to remove redundancy, producing a compact evidence set with lightweight location IDs. For generation, these IDs act as precise grounding signals that guide the LLM to the correct memory positions, mitigating the lost-in-the-middle effect while preserving original contextual integrity. Extensive experiments on four benchmarks demonstrate that MemLoc achieves state-of-the-art retrieval accuracy and response quality while maintaining efficiency. Our code is available at: https://github.com/Nikol-coder/MemLoc.