Research

Paper

AI LLM February 24, 2026

Autonomous Laboratory Agent via Customized Domain-Specific Language Model and Modular AI Interface

Authors

Zhuo Diao, Kouma Matsumoto, Linfeng Hou, Hayato Yamashita, Masayuki Abe

Abstract

We introduce a system architecture that addresses a fundamental challenge in deploying language-model agents for autonomous control of scientific instrumentation: ensuring reliability in safety-critical environments. The framework integrates probabilistic reasoning by domain-specialized language models with deterministic execution layers that enforce constraints through structured validation and modular orchestration. By separating intent interpretation, experimental planning, and command verification, the architecture translates high-level scientific goals into verifiable experimental actions. We demonstrate this approach in real-time atomic-resolution scanning probe microscopy experiments operated at room temperature, where the system autonomously generates control strategies, invokes corrective modules, and maintains stable operation under experimentally challenging conditions. Quantitative evaluations show that domain-adapted small language models achieve high routing robustness and command accuracy while operating on consumer-grade hardware. Beyond a specific instrument, the framework establishes a general computational principle for deploying language-model agents in safety-critical experimental workflows, providing a pathway toward scalable autonomous laboratories.

Metadata

arXiv ID: 2602.20669
Provider: ARXIV
Primary Category: physics.app-ph
Published: 2026-02-24
Fetched: 2026-02-25 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.20669v1</id>\n    <title>Autonomous Laboratory Agent via Customized Domain-Specific Language Model and Modular AI Interface</title>\n    <updated>2026-02-24T08:20:00Z</updated>\n    <link href='https://arxiv.org/abs/2602.20669v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.20669v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>We introduce a system architecture that addresses a fundamental challenge in deploying language-model agents for autonomous control of scientific instrumentation: ensuring reliability in safety-critical environments. The framework integrates probabilistic reasoning by domain-specialized language models with deterministic execution layers that enforce constraints through structured validation and modular orchestration. By separating intent interpretation, experimental planning, and command verification, the architecture translates high-level scientific goals into verifiable experimental actions. We demonstrate this approach in real-time atomic-resolution scanning probe microscopy experiments operated at room temperature, where the system autonomously generates control strategies, invokes corrective modules, and maintains stable operation under experimentally challenging conditions. Quantitative evaluations show that domain-adapted small language models achieve high routing robustness and command accuracy while operating on consumer-grade hardware. Beyond a specific instrument, the framework establishes a general computational principle for deploying language-model agents in safety-critical experimental workflows, providing a pathway toward scalable autonomous laboratories.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='physics.app-ph'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cond-mat.mtrl-sci'/>\n    <published>2026-02-24T08:20:00Z</published>\n    <arxiv:primary_category term='physics.app-ph'/>\n    <author>\n      <name>Zhuo Diao</name>\n    </author>\n    <author>\n      <name>Kouma Matsumoto</name>\n    </author>\n    <author>\n      <name>Linfeng Hou</name>\n    </author>\n    <author>\n      <name>Hayato Yamashita</name>\n    </author>\n    <author>\n      <name>Masayuki Abe</name>\n    </author>\n  </entry>"
}