Accurate recall from large scale memories remains a core challenge for memory augmented AI assistants performing question answering (QA), especially in similarity dense scenarios where existing methods mainly rely on semantic distance to the query for retrieval. Inspired by how humans link information associatively, we propose AssoMem, a novel framework constructing an associative memory graph that anchors dialogue utterances to automatically extracted clues. This structure provides a rich organizational view of the conversational context and facilitates importance aware ranking. Further, AssoMem integrates multi-dimensional retrieval signals-relevance, importance, and temporal alignment using an adaptive mutual information (MI) driven fusion strategy. Extensive experiments across three benchmarks and a newly introduced dataset, MeetingQA, demonstrate that AssoMem consistently outperforms SOTA baselines, verifying its superiority in context-aware memory recall.
@article{arxiv.2510.10397,
title = {AssoMem: Scalable Memory QA with Multi-Signal Associative Retrieval},
author = {Kai Zhang and Xinyuan Zhang and Ejaz Ahmed and Hongda Jiang and Caleb Kumar and Kai Sun and Zhaojiang Lin and Sanat Sharma and Shereen Oraby and Aaron Colak and Ahmed Aly and Anuj Kumar and Xiaozhong Liu and Xin Luna Dong},
journal= {arXiv preprint arXiv:2510.10397},
year = {2025}
}