When Does Overlap Help? OSU-Mem and a Cell-Conditional Analysis of Trajectory Memory for LLM Agents

• URL: http://arxiv.org/abs/2606.28376v1
• Date: Fri, 19 Jun 2026 04:23:04 GMT
• Title: When Does Overlap Help? OSU-Mem and a Cell-Conditional Analysis of Trajectory Memory for LLM Agents
• Abstract summary: Long-horizon large language model (LLM) agents accumulate interaction trajectories that quickly exceed any practical prompt budget.<n>We ask a mechanism question: when does organizing trajectory memory into overlapping semantic units (OSUs) help retrieval over flat or disjoint alternatives?<n>We instantiate this in OSUMem, which retrieves from an overlapping OSU pool via budgeted coarse-to-fine expansion.
• Score: 8.361893760388329
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Long-horizon large language model (LLM) agents accumulate interaction trajectories that quickly exceed any practical prompt budget, and existing memory methods either truncate aggressively and lose non-local evidence or retain boilerplate that degrades decision quality. We ask a mechanism question rather than claiming a better general-purpose memory system: when does organizing trajectory memory into overlapping semantic units (OSUs) -- groups of related steps in which one step may belong to several units -- help retrieval over flat or disjoint alternatives? We instantiate this in OSU-Mem, which retrieves from an overlapping OSU pool via budgeted coarse-to-fine expansion, and show its benefit is conditional: overlapping memory helps when the evidence steps a query needs share tool calls or entities, but hurts when those steps are fully heterogeneous and share neither. On a synthetic benchmark where evidence carries such shared structure by construction, OSU-Mem improves over the strongest baseline as the theory predicts; yet on a concatenated, constructed unaugmented $τ$-bench setting its aggregate advantage over flat retrieval vanishes. Splitting queries by whether their evidence shares tools and entities shows this near-tie to be an artifact of mixing query types rather than a property of either method, and ToolBench, a controlled probe built to carry shared structure by design, corroborates the same mechanism via an overlap-vs.-disjoint construction contrast (under a coverage-guided variant), isolating the construction principle rather than validating the full default system. Because the relevant sharing is cheaply estimable from metadata, the analysis yields a metadata-based heuristic for predicting when overlap is likely to improve retrieval. We deliberately isolate the retrieval layer, assessed by retrieval quality and an LLM-mediated evidence-selection stage.

Related papers

• GraphMemix: Query-Aware Evidence Forests for Long-Term Multimodal Agent Memory [87.58855368314693]
  We propose GraphMemix, a graph memory framework that models memory organization as query-aware evidence-forest construction.<n> Experimental results across four long-term multimodal memory benchmarks demonstrate significant improvements with different foundation models.
  arXiv  Detail & Related papers  (2026-08-27T11:28:49Z)
• Integration-First Structural Coverage for Embedded Software:Trace-Based Evidence, Hybrid Runtime Analysis, and Cross-Variant Consolidation [42.52742840006677]
  We describe an integrationfirst coverage strategy that treats integration and system tests as the baseline measurement.<n>We also describe Hyper Coverage, a consolidation layer that merges evidence across test levels, test runs, variants and build configurations.
  arXiv  Detail & Related papers  (2026-08-13T14:46:33Z)
• Lightweight Chunk Selection for Mobile Retrieval-Augmented Generation [8.227745781139634]
  We study lightweight RAG chunk selection as an evidence-alignment problem.<n>Our selector combines three complementary feature sources: question hidden states that represent LLM-side query intent, MoE routing-derived expert signals that capture the generator's internal routing structure, and retrieved chunk embeddings that preserve candidate-side evidence geometry.<n>Experiments show that the proposed selector consistently improves rank-1 evidence selection over mobile-applicable baselines by an average of 2.5%.
  arXiv  Detail & Related papers  (2026-08-04T05:26:25Z)
• Imputation Meets Clustering: Exploiting Latent Subgroup Structure for Missing Data Recovery [52.29866804948133]
  We propose CAGI (Cluster-Aware Generative Imputation), a framework that reformulates clustering and imputation as a mutually reinforcing co-optimization process.<n>Experiments on 14 benchmark datasets with 15 representative baselines demonstrate the superiority of CAGI.
  arXiv  Detail & Related papers  (2026-07-08T02:50:02Z)
• Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents [51.30250860677378]
  We propose MRAgent, a framework that combines an associative memory graph with an active reconstruction mechanism.<n>Running on a Cue-Tag-Content graph, our active reconstruction mechanism integrates reasoning directly into memory access.<n>Experiments on the LoCoMo benchmark and LongMemEval benchmark demonstrate significant improvements over strong baselines.
  arXiv  Detail & Related papers  (2026-06-04T11:29:46Z)
• eMEM: A Hybrid Spatio-Temporal Memory System For Embodied Agents [1.8047694351309203]
  We present eMEM, a hybrid graph-based memory system for embodied agents operating in physical environments.<n>eMEM fills this gap with a multi-index architecture (ITE for structured storage, hlib for approximate nearest neighbour semantic search, and an R-tree for spatial queries)<n>We also introduce eMEM-Bench v1, a benchmark we construct over ProcTHOR-10K scenes for embodied memory evaluation.
  arXiv  Detail & Related papers  (2026-06-02T09:22:12Z)
• EfficientGraph-RAG: Structured Retrieval-State Management for Cross-Task Retrieval-Augmented Generation [45.66803833786235]
  Retrieval-augmented generation has become the standard way to ground large language models in external knowledge.<n> EfficientGraph-RAG ranks first on the reported answer-quality metrics averaged over the three evaluated LongBench retrieval-style subsets.
  arXiv  Detail & Related papers  (2026-05-25T03:08:53Z)
• Agentic DAG-Orchestrated Planner Framework for Multi-Modal, Multi-Hop Question Answering in Hybrid Data Lakes [5.276278723196607]
  Agentic DAG-Orchestrated Transformer (A.DOT) Planner is a framework for multi-modal, multi-hop question answering.<n>A.DOT compiles user NL queries into directed acyclic graph (DAG) execution plans spanning both structured and unstructured stores.<n>System decomposes queries into parallelizable sub-queries, incorporates schema-aware reasoning, and applies both structural and semantic validation.
  arXiv  Detail & Related papers  (2026-03-15T05:34:16Z)
• Diagnosing Structural Failures in LLM-Based Evidence Extraction for Meta-Analysis [0.8193467416247519]
  Review and meta-analyses rely on converting narrative articles into structured, numerically grounded study records.<n>Despite rapid advances in large language models (LLMs), it remains unclear whether they can meet the structural requirements of this process.<n>We propose a structural, diagnostic framework that evaluates LLM-based evidence extraction as a progression of schema-constrained queries.
  arXiv  Detail & Related papers  (2026-02-11T14:09:43Z)
• Beyond RAG for Agent Memory: Retrieval by Decoupling and Aggregation [22.803751188961865]
  We argue retrieval should move beyond similarity matching and instead operate over latent components.<n>We propose xMemory, which builds a hierarchy of intact units and maintains a searchable high-level node organisation.
  arXiv  Detail & Related papers  (2026-02-02T12:04:58Z)
• AMA: Adaptive Memory via Multi-Agent Collaboration [54.490349689939166]
  We propose Adaptive Memory via Multi-Agent Collaboration (AMA), a novel framework that leverages coordinated agents to manage memory across multiple granularities.<n>AMA significantly outperforms state-of-the-art baselines while reducing token consumption by approximately 80% compared to full-context methods.
  arXiv  Detail & Related papers  (2026-01-28T08:09:49Z)
• SiReRAG: Indexing Similar and Related Information for Multihop Reasoning [96.60045548116584]
  SiReRAG is a novel RAG indexing approach that explicitly considers both similar and related information.<n>SiReRAG consistently outperforms state-of-the-art indexing methods on three multihop datasets.
  arXiv  Detail & Related papers  (2024-12-09T04:56:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.