Preprint

Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses

Zhaochen Yu Yingcheng Wu Zhen-Fei Yin Kaiyuan Chen Zhe Zhao Mengdi Wang Shuicheng Yan Ling Yang

Aug 2026 · 0 citations · 50 references

Computer Science

TL;DR

Recuris, a recursive Experiential-Working Memory architecture for long-horizon agent harnesses, in which Working Memory tracks task progress and guides skill selection from Experiential Memory, grounding skill use in current needs rather than the full history, positions recursively evolving memory as a scalable foundation for RSI.

Abstract

Recursive self-improvement (RSI) remains hard in long-horizon tasks, where growing histories obscure the task state and misalign skill invocation. We introduce Recuris, a recursive Experiential-Working Memory architecture for long-horizon agent harnesses, in which Working Memory tracks task progress and guides skill selection from Experiential Memory, grounding skill use in current needs rather than the full history. This coupling also turns execution into structured evidence that localizes failures to specific memory components. Across tasks, a fixed Meta-Agent turns that evidence into localized, validation-gated updates to Skill Memory that reshape execution and yield new evidence, forming a bounded recursive memory-evolution loop. Across four long-horizon benchmarks and ten models, Recuris improves task success in 35 of the 37 completed model-benchmark pairs, carrying frontier models to SOTA-level task success: on tau-bench it adds +17.8 points to GPT-5.6 Sol and +15.6 to Claude Opus 5, taking Opus 5 to 87.9%, and +16.6/+13.5 points on Qwen3.6-27B/35B on SkillFlow. The advantage widens as the interaction horizon grows, to +32.2 points on the longest tasks, and common long-horizon failures fall by up to 80%. These results position recursively evolving memory as a scalable foundation for RSI, enabling agents to continuously transform accumulated experience into increasingly effective long-horizon behavior. Code: https://github.com/Gen-Verse/Recuris

arXiv Semantic Scholar

View source

Similar papers

Preprint Aug 2026

MemPrism: Task-Conditioned Relational Memory Views for Long-Horizon Agents

Experiments on long-horizon embodied and web-agent benchmarks show that MemPrism consistently improves the task performance, especially as trajectories become longer, while reducing memory token consumption.

Zhi-Sheng Chen, Bingfan Zeng, Bangde Cao et al. · 0 citations

Review Aug 2026

The Horizon Gap: Planning, Memory, Execution, Training, and Evaluation for Long-Horizon LLM Agents

A corpus of 1,547 arXiv papers collected via systematic seed harvest with a disclosed 26.8% bleed filter is surveyed, finding the same pattern: outcome-only signals grow uninformative as horizons lengthen, and the field's response manufactures denser step-level signals.

Mingguang Chen, Licheng Wang, Bo Qu · 0 citations

Jul 2026

PRO-LONG: Programmatic Memory Enables Long-Horizon Reasoning

Pro-LONG, a minimal context management framework built around programmatic memory for LLM agents in long-horizon, exploratory settings, is proposed, which addresses the tradeoff of preserving more information makes retrieving relevant details less tractable.

A. Fox, Jun-Lin Wang, P. Rosu et al. · arXiv.org · 2 citations · ⚡1

#artificial intelligence Preprint Sep 2026

CHIME: Credit-Aware Hierarchical Memory Evolution for Long-Horizon Agentic Planning

Planning is a central capability that enables agents to decompose complex long-horizon tasks into manageable steps. Test-time search and training-based methods improve planning but incur high inference costs or require expensive training data. Self-evolving memory instead accumulates reusable experience from agent interaction outcomes into an external memory bank, so planning capability keeps improving at inference time without parameter updates. However, existing self-evolving memory methods share an inherent credit assignment problem: they rely on final task outcomes as feedback, but such outcomes conflate plan quality with execution errors and environmental factors, so the accumulated planning experience is often biased and noisy. To address this problem, we propose Credit-Aware Hierarchical Memory Evolution (CHIME), a self-evolving memory framework that maintains a separate planning bank and execution bank and follows an attribute-before-memorize principle: CHIME first attributes each task outcome to the plan, the execution, both, or neither, and then updates only the corresponding memory bank. Extensive experiments on four long-horizon agent benchmarks show that CHIME consistently outperforms state-of-the-art training-based and self-evolving memory baselines. Further analyses reveal several interesting findings. For example, CHIME accumulates effective memory with far fewer items. In addition, the learned memory values faithfully reflect downstream utility: high-quality planning memories are more valuable than execution memories. Finally, the accumulated memory effectively transfers across backbone models. Code will be released at https://github.com/ATH-MaaS/Marco-DeepResearch.

Yongshi Ye, Tian Lan, Feihu Jiang et al. · 0 citations

Preprint Aug 2026

MemArbiter: Decision-Time Memory Arbitration for Long-Horizon LLM Agents

Results show that function-aware memory arbitration enables accessible information to guide actions more effectively, and improves post-failure recovery and reduces failed-action repetition and state-action recurrence.

Jiajun Dong, Yutao Hu, Fengrui Fan et al. · 0 citations

Jul 2026

From Memory to Skills: Evidence-Grounded Co-Evolution Governance for Long-Horizon LLM Agents

This paper proposes MSCE, a training-free Memory--Skill Co-Evolution framework that organizes agent experience into grounded step traces, reusable procedural policies, and declarative environmental cognition, and introduces reflection-weighted value backfilling.

Bo Tang, Yang Zhang, Guomian Zhuang et al. · arXiv.org · 1 citation