MemFuse introduces a multi‑source memory fusion system and a corresponding benchmark, MemFuseBench, designed to evaluate agents that must integrate fragmented observations from multiple applications, devices, users, and time points. The benchmark uses a Scene‑to‑Sensor pipeline to generate source‑tagged observations, evidence‑grounded questions, and adversarial distractors, enabling systematic assessment of temporal reasoning, cross‑source evidence fusion, and noise robustness. Experiments show that MemFuse outperforms other memory systems across all evaluated LLM settings, especially on questions requiring cross‑source evidence fusion.
MemoryLACE (MemLACE) is a lightweight memory framework that explicitly models the lifecycle of textual evidence—capturing sparse merge, supersession, and contradiction relations—while preserving atomic natural‑language memories and their provenance. Unlike traditional systems that retrieve memories independently, MemLACE reconstructs relation‑aware evidence units that expose current, historical, supporting, and conflicting evidence for downstream reasoning. In benchmark evaluations (BEAM and StructMemEval) using both open‑weight and proprietary LLM backbones, MemLACE achieves the highest overall performance among same‑backbone comparisons and reduces BEAM runtime by 66.6% compared to the strongest reflective‑memory baseline, Hindsight.
By Meriem Yacoubi, Pia Schmidt, Nenad Petrovic, Ahmed Frikha, Martin Kirchhoff, Alois Knoll
UTILMEM is a new diagnostic benchmark that tests how conversational agents use long‑term memory, focusing on reasoning over dense histories, spotting implicitly relevant memories, synthesizing distributed evidence, and resisting interference from similar distractors. It contains 1,717 instances across five domains and evaluates a range of retrieval‑based and memory‑augmented systems. The study shows that strong performance on traditional factual recall does not guarantee effective memory utilization, highlighting a gap between retrieving information and integrating it into coherent, task‑oriented outputs.
By Peijun Qing, Fobo Shi, Soroush Vosoughi
arXiv:2605. 28831v2 Announce Type: replace-cross Abstract: Long-horizon memory question answering often requires sparse evidence from heterogeneous histories, including events, object states, visual observations, temporal relations, and causal steps.
By Encheng Su, Jianyu Wu, Jinouwen Zhang, Qiucheng Yu, Chen Tang, Pengze Li, Lintao Wang, Aoran Wang, Xinzhu Ma, Shixiang Tang, Yizhou Wang, Houqiang Li
arXiv:2606. 28349v1 Announce Type: cross Abstract: Long-context reasoning requires models to access, retrieve, and integrate evidence scattered across documents, dialogues, and accumulated interaction histories.
By Zeju Li, Ziyang Zheng, Yizhou Zhou, Qiang Xu
arXiv:2608. 03463v1 Announce Type: new Abstract: Long-term memory is essential for LLM-based agents to sustain interactions and reliably leverage distant history.
By Yuxin Liao, Le Wu, Min Hou, Hao Liu, Han Wu, Zishu Wang
arXiv:2609.00551v1 Announce Type: cross
Abstract: Multimodal memory offers a scalable interface for long-video question answering, but existing methods often retrieve captions, frames, transcripts, s...
By Yijun Chen, Yaqi Zheng, Yanya Li, Boyi Xiao, Buqiang Xu, Shuofei Qiao, Jizhan Fang, Xinle Deng, Yunzhi Yao, Xuehai Wang, Liuxin Zhang, Hui Li, Huajun Chen, Shumin Deng
MemCoRe is a memory system for large language model agents that organizes factual knowledge into a compression hierarchy, progressively reducing redundancy while preserving retrieval structure. The hierarchy compresses detailed records into keywords and then into topic groups, allowing evidence to be located by searching across levels. Experiments show that MemCoRe outperforms current state‑of‑the‑art baselines in retrieving relevant evidence for downstream reasoning.
By Zhenyuan Zhang, Xianzhang Jia, Zhiqin Yang, Zhenbo Song, Wei Xue, Sirui Han, Yike Guo
The paper introduces SCALE-QA, a new QA benchmark that tests conversational memory in flat, unsegmented multi‑topic threads by requiring agents to infer which earlier episode supports a later task decision. The dataset contains 3,000 audited questions across ten domains, uses deterministic four‑way multiple‑choice grading, and includes a runtime builder for reproducibility. The authors also propose Temporal‑Semantic Interleaved Memory Reconstruction (TSIM), a hierarchical memory stack that segments turns into coherent episodes and indexes them with deterministic summaries and cluster‑routing views, achieving significant accuracy gains over strong RAG baselines and long‑context LLMs.
By Zhexi Feng, Ruiyi Zhang, Yongbo Yang, Pengtao Xie
arXiv:2601. 21714v5 Announce Type: replace Abstract: The evolution of Large Language Model (LLM) agents towards System~2 reasoning, characterized by deliberative, high-precision problem-solving, requires maintaining rigorous logical integrity over extended horizons.
By Kaixiang Wang, Yidan Lin, Jiong Lou, Zhaojiacheng Zhou, Bunyod Suvonov, Jie Li
The paper investigates how large language model agents can over-rely on agentic memory, a phenomenon where retrieved memories distort inference even when they are correctly stored and retrieved. It shows that memory is helpful when past experience fully transfers to the current task but becomes misleading under partial query-memory overlap, a pattern confirmed by controlled experiments. To address this, the authors propose MEMTRIM, a plug‑and‑play framework that indexes memory evidence at write time and limits its reuse at read time, thereby reducing over-reliance without retraining and preserving useful memory benefits across models and memory architectures.
By Luoxi Tang, Yuqiao Meng, Nilesh Auradkar, Muchao Ye, Dazheng Zhang, Zhaohan Xi
The paper introduces MemCompile, a query‑conditioned evidence compilation framework designed to improve long‑term conversational memory. It uses an Address Plane to locate relevant support via compact descriptors and capacity‑aware cells, then normalizes retrieval hits into stable SourcePacket identities. An Evidence Plane allocates a bounded evidence budget across packets, admits limited additional support, and produces answer‑ready context, achieving 80.52% Judge Accuracy on LoCoMo and improving the quality‑cost Pareto frontier.
By Shangze Li, Yanyu Yao, Derong Xu, Zhi Zheng, Tong Xu