When Does Memory Help Multi-Trajectory Inference for Tool-Use LLM Agents?

• URL: http://arxiv.org/abs/2605.28224v1
• Date: Wed, 27 May 2026 09:39:19 GMT
• Title: When Does Memory Help Multi-Trajectory Inference for Tool-Use LLM Agents?
• Abstract summary: Multi-trajectory inference for tool-use LLM agents generates multiple reasoning attempts and selecting among them.<n>Existing cross-trajectory memory methods are each evaluated under a single inference strategy on a single task.<n>We propose a unified framework that decomposes memory along two axes -- the scope of transfer and the abstraction of the transferred content.
• Score: 3.48949373776636
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multi-trajectory inference for tool-use LLM agents - generating multiple reasoning attempts and selecting among them - benefits from transferring knowledge across attempts so that later ones avoid the pitfalls of earlier ones. Existing cross-trajectory memory methods (trajectory-level reflection, atomic fact extraction, raw observation injection) are each evaluated under a single inference strategy on a single task, making it unclear whether reported gains reflect properties of the memory abstraction or of the inference method. We propose a unified framework that decomposes memory along two axes -- the scope of transfer (within an expansion vs. across trajectories) and the abstraction of the transferred content -- and evaluate four methods under three inference strategies (best-of-N, beam search, MCTS) on four tool-use benchmarks spanning SQL, knowledge-graph, and CLI environments, in a verifier-free setting that matches the deployment regime of practical agents. The experiment matrix identifies the inference method as a confound: the same memory method produces statistically distinct results under different inference strategies on the same examples. Reflection reaches significance only under MCTS (not under best-of-N); within-expansion injection (conditioning each candidate on prior siblings' outcomes) helps only diversity-starved beam search; and atomic fact extraction is accuracy-neutral but shortens trajectories by 19-26% on tasks with reusable environmental structure.

Related papers

• Learning to Retrieve: Dual-Level Long-Term Memory for Text-to-SQL Agents [63.38073530162493]
  Long-term memory helps agents reuse past experiences.<n> Static methods rely on fixed similaritys that do not optimize downstream utility.<n>We propose MERIT, a dynamic multi-horizon memory retrieval framework.
  arXiv  Detail & Related papers  (2026-05-30T05:44:34Z)
• Memory in the LLM Era: Modular Architectures and Strategies in a Unified Framework [21.68420785280451]
  Memory emerges as the core module in the large language model (LLM)-based agents for long-horizon complex tasks.<n>We first summarize a unified framework that incorporates all the existing agent memory methods from a high-level perspective.<n>We then extensively compare representative agent memory methods on two well-known benchmarks and examine the effectiveness of all methods.
  arXiv  Detail & Related papers  (2026-04-02T07:19:20Z)
• TopoCurate:Modeling Interaction Topology for Tool-Use Agent Training [53.93696896939915]
  Training tool-use agents typically rely on Supervised Fine-Tuning (SFT) on successful trajectories and Reinforcement Learning (RL) on pass-rate-selected tasks.<n>We propose TopoCurate, an interaction-aware framework that projects multi-trial rollouts from the same task into a unified semantic quotient topology.<n>TopoCurate achieves consistent gains of 4.2% (SFT) and 6.9% (RL) over state-of-the-art baselines.
  arXiv  Detail & Related papers  (2026-03-02T10:38:54Z)
• A Benchmark for Procedural Memory Retrieval in Language Agents [0.023227405857540805]
  Current AI agents excel in familiar settings, but fail sharply when faced with novel tasks with unseenProc.<n>We present the first benchmark that isolates procedural memory retrieval from task execution.<n>Our results expose a clear generalization cliff: embedding-based methods perform strongly on familiar contexts, yet degrade considerably on novel ones.
  arXiv  Detail & Related papers  (2025-11-21T08:08:53Z)
• What Works for 'Lost-in-the-Middle' in LLMs? A Study on GM-Extract and Mitigations [1.2879523047871226]
  GM-Extract is a novel benchmark dataset meticulously designed to evaluate LLM performance on retrieval of control variables.<n>We conduct a systematic evaluation of 7-8B parameter models on two multi-document tasks (key-value extraction and question-answering)<n>While a distinct U-shaped curve was not consistently observed, our analysis reveals a clear pattern of performance across models.
  arXiv  Detail & Related papers  (2025-11-17T20:50:50Z)
• Task-Specific Data Selection for Instruction Tuning via Monosemantic Neuronal Activations [19.25205110583291]
  A critical bottleneck is selecting the most relevant data to maximize task-specific performance.<n>Existing data selection approaches include unstable influence-based methods and more stable distribution alignment methods.<n>We introduce a dedicated similarity metric for this space to better identify task-relevant data.
  arXiv  Detail & Related papers  (2025-03-19T11:35:57Z)
• Collaborative Feature-Logits Contrastive Learning for Open-Set Semi-Supervised Object Detection [75.02249869573994]
  In open-set scenarios, the unlabeled dataset contains both in-distribution (ID) classes and out-of-distribution (OOD) classes.<n>Applying semi-supervised detectors in such settings can lead to misclassifying OOD class as ID classes.<n>We propose a simple yet effective method, termed Collaborative Feature-Logits Detector (CFL-Detector)
  arXiv  Detail & Related papers  (2024-11-20T02:57:35Z)
• A Unified Approach Towards Active Learning and Out-of-Distribution Detection [49.334006405481375]
  We introduce SISOM as the first unified solution for both AL and OOD detection.<n>By leveraging feature space distance metrics SISOM combines the strengths of the currently independent tasks to solve both effectively.<n>In AL, SISOM outperforms others and delivers top-1 performance in three benchmarks.
  arXiv  Detail & Related papers  (2024-05-18T17:02:57Z)
• Spatial Cascaded Clustering and Weighted Memory for Unsupervised Person Re-identification [32.95715593278961]
  Unsupervised person re-identification (re-ID) methods achieve high performance by leveraging fine-grained local context. Part-based methods obtain local contexts through horizontal division, which suffer from misalignment due to various human poses. We introduce the Spatial Cascaded Clustering and Weighted Memory (SCWM) method to address these challenges. SCWM aims to parse and align more accurate local contexts for different human body parts while allowing the memory module to balance hard example mining and noise suppression.
  arXiv  Detail & Related papers  (2024-03-01T03:52:29Z)
• 3D-QueryIS: A Query-based Framework for 3D Instance Segmentation [74.6998931386331]
  Previous methods for 3D instance segmentation often maintain inter-task dependencies and the tendency towards a lack of robustness. We propose a novel query-based method, termed as 3D-QueryIS, which is detector-free, semantic segmentation-free, and cluster-free. Our 3D-QueryIS is free from the accumulated errors caused by the inter-task dependencies.
  arXiv  Detail & Related papers  (2022-11-17T07:04:53Z)
• Plug-and-Play Few-shot Object Detection with Meta Strategy and Explicit Localization Inference [78.41932738265345]
  This paper proposes a plug detector that can accurately detect the objects of novel categories without fine-tuning process. We introduce two explicit inferences into the localization process to reduce its dependence on annotated data. It shows a significant lead in both efficiency, precision, and recall under varied evaluation protocols.
  arXiv  Detail & Related papers  (2021-10-26T03:09:57Z)

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.