English

Experience-Driven Multi-Agent Systems Are Training-free Context-aware Earth Observers

Artificial Intelligence 2026-02-04 v1 Computer Vision and Pattern Recognition Machine Learning Multiagent Systems

Abstract

Recent advances have enabled large language model (LLM) agents to solve complex tasks by orchestrating external tools. However, these agents often struggle in specialized, tool-intensive domains that demand long-horizon execution, tight coordination across modalities, and strict adherence to implicit tool constraints. Earth Observation (EO) tasks exemplify this challenge due to the multi-modal and multi-temporal data inputs, as well as the requirements of geo-knowledge constraints (spectrum library, spatial reasoning, etc): many high-level plans can be derailed by subtle execution errors that propagate through a pipeline and invalidate final results. A core difficulty is that existing agents lack a mechanism to learn fine-grained, tool-level expertise from interaction. Without such expertise, they cannot reliably configure tool parameters or recover from mid-execution failures, limiting their effectiveness in complex EO workflows. To address this, we introduce \textbf{GeoEvolver}, a self-evolving multi-agent system~(MAS) that enables LLM agents to acquire EO expertise through structured interaction without any parameter updates. GeoEvolver decomposes each query into independent sub-goals via a retrieval-augmented multi-agent orchestrator, then explores diverse tool-parameter configurations at the sub-goal level. Successful patterns and root-cause attribution from failures are then distilled in an evolving memory bank that provides in-context demonstrations for future queries. Experiments on three tool-integrated EO benchmarks show that GeoEvolver consistently improves end-to-end task success, with an average gain of 12\% across multiple LLM backbones, demonstrating that EO expertise can emerge progressively from efficient, fine-grained interactions with the environment.

Keywords

Cite

@article{arxiv.2602.02559,
  title  = {Experience-Driven Multi-Agent Systems Are Training-free Context-aware Earth Observers},
  author = {Pengyu Dai and Weihao Xuan and Junjue Wang and Hongruixuan Chen and Jian Song and Yafei Ou and Naoto Yokoya},
  journal= {arXiv preprint arXiv:2602.02559},
  year   = {2026}
}

Comments

21 pages, 6 figures

R2 v1 2026-07-01T09:32:39.905Z