Agents for self-driving laboratories applied to quantum computing

2024-12-10

arXiv.org

AI Summary

Unstructured, multimodal knowledge in quantum experiments poses significant challenges for integration and automated utilization. Method: This paper introduces k-agents—a framework leveraging large language models (LLMs) to encapsulate laboratory knowledge, employing a state-machine-driven execution agent to model multi-step experimental workflows, and enabling closed-loop decision-making via multi-agent collaborative reasoning and real-time experimental feedback. Contribution/Results: k-agents pioneers a knowledge-encapsulation–based multi-agent architecture. It achieves, for the first time on superconducting quantum processors, fully autonomous calibration and high-fidelity entangled-state preparation—completing experiment planning, execution, analysis, and characterization continuously over several hours. Performance matches that of human experts, markedly enhancing scientific discovery efficiency and experimental reproducibility.

Technology Category

• Machine Learning: Quantum Machine Learning
• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Fully automated self-driving laboratories are promising to enable high-throughput and large-scale scientific discovery by reducing repetitive labour. However, effective automation requires deep integration of laboratory knowledge, which is often unstructured, multimodal, and difficult to incorporate into current AI systems. This paper introduces the k-agents framework, designed to support experimentalists in organizing laboratory knowledge and automating experiments with agents. Our framework employs large language model-based agents to encapsulate laboratory knowledge including available laboratory operations and methods for analyzing experiment results. To automate experiments, we introduce execution agents that break multi-step experimental procedures into state machines, interact with other agents to execute each step and analyze the experiment results. The analyzed results are then utilized to drive state transitions, enabling closed-loop feedback control. To demonstrate its capabilities, we applied the agents to calibrate and operate a superconducting quantum processor, where they autonomously planned and executed experiments for hours, successfully producing and characterizing entangled quantum states at the level achieved by human scientists. Our knowledge-based agent system opens up new possibilities for managing laboratory knowledge and accelerating scientific discovery.

Problem

• Automating quantum experiments with AI-driven agents
• Organizing unstructured lab knowledge for AI integration
• Enabling closed-loop control in self-driving laboratories

Innovation

• LLM-based agents encapsulate lab knowledge
• Execution agents automate multi-step experiments
• Closed-loop feedback controls state transitions