Abstract:The successful integration of large language models (LLMs) into laboratory workflows has demonstrated robust capabilities in natural language processing, autonomous task execution, and collaborative problem-solving. This offers an exciting opportunity to realize the dream of autonomous chemical research on demand. Here, we report a robotic AI chemist powered by a hierarchical multi-agent system, ChemAgents, based on an on-board Llama-3-70B LLM, capable of executing complex, multi-step experiments with minimal human intervention. It operates through a Task Manager agent that interacts with human researchers and coordinates four role-specific agents—Literature Reader, Experiment Designer, Computation Performer, and Robot Operator—each leveraging one of four foundational resources: a comprehensive Literature Database, an extensive Protocol Library, a versatile Model Library, and a state-of-the-art Automated Lab. We demonstrate its versatility and efficacy through six experimental tasks of varying complexity, ranging from straightforward synthesis and characterization to more complex exploration and screening of experimental parameters, culminating in the discovery and optimization of functional materials. Our multi-agent-driven robotic AI chemist showcases the potential of on-demand autonomous chemical research to drive unprecedented efficiencies, accelerate discovery, and democratize access to advanced experimental capabilities across academic disciplines and industries.

What problem does this paper attempt to address?

The problem this paper attempts to address is: How to utilize large language models (LLMs) and multi-agent systems to construct a robot AI chemist capable of autonomously performing complex chemical experiments, thereby achieving on-demand autonomous chemical research. Specifically, the paper proposes a hierarchical multi-agent system-based robot AI chemist (ChemAgents) that can execute multi-step, multi-site, and/or multi-robot complex experimental tasks with minimal human intervention. ### Main Issues: 1. **Automation of complex experimental tasks**: Traditional laboratory automation systems have limitations in handling complex, multi-step experimental tasks, especially when multiple sites or multiple robots need to work collaboratively. The method proposed in the paper aims to overcome these limitations and achieve efficient, autonomous experimental execution. 2. **Collaboration and coordination of multi-agent systems**: To effectively handle complex experimental tasks, collaboration and coordination among multiple agents (such as literature readers, experiment designers, computation performers, and robot operators) are required. The paper designs a central Task Manager to coordinate the work of these agents. 3. **Automated planning and management of experimental tasks**: The paper proposes a system that can generate experimental plans from the demand descriptions of human researchers and manage the entire experimental process, including experiment design, robot operation, and data processing. 4. **Analysis and optimization of experimental data**: For research tasks requiring data-driven discovery or closed-loop iterative optimization, the paper introduces a Computation Performer agent that uses pre-trained machine learning models and Bayesian optimization techniques to improve experimental efficiency and result quality. ### Solutions: - **Multi-agent system (ChemAgents)**: Composed of a Task Manager and four role-specific agents, each responsible for different tasks such as literature search, experiment design, computation execution, and robot operation. - **Basic resource library**: Includes a literature database, experiment protocol library, machine learning model library, and automated laboratory, providing necessary tools and information support for the agents. - **Hierarchical task decomposition and coordination**: The Task Manager decomposes complex experimental tasks into multiple sub-tasks and coordinates the work of each agent to ensure the smooth progress of the experimental process. - **Autonomous experimental execution**: The robot operator agent translates the experiment design into robot-executable code, achieving automated execution of the experiment. In summary, this paper addresses the automation of complex chemical experiments by constructing a multi-level, multi-agent robot AI chemist system, providing new possibilities for future chemical research.

A multi-agent-driven robotic AI chemist enabling autonomous chemical research on demand

Multi-agent based distributed chemical process monitoring and diagnosis

An Autonomous Large Language Model Agent for Chemical Literature Data Mining

An all-round AI-Chemist with a scientific mind

Chemist-X: Large Language Model-empowered Agent for Reaction Condition Recommendation in Chemical Synthesis

Accelerated end-to-end chemical synthesis development with large language models

DrugAgent: Automating AI-aided Drug Discovery Programming through LLM Multi-Agent Collaboration

AI-chemist for Chemistry Synthesis, Property Characterization, and Performance Testing

Augmenting large language models with chemistry tools

Towards AI Research Agents in the Chemical Sciences

A multi-robot-multi-task scheduling system for autonomous chemistry laboratories

Autonomous mobile robots for exploratory synthetic chemistry

A robotic AI-Chemist system for multi-modal AI-ready database

A Review of Large Language Models and Autonomous Agents in Chemistry

ARChemist: Autonomous Robotic Chemistry System Architecture

AIR-Chem: Authentic Intelligent Robotics for Chemistry

Autonomous chemical science and engineering enabled by self-driving laboratories

ChemCrow: Augmenting large-language models with chemistry tools

Autonomous chemical research with large language models

AI-driven robotic chemist for autonomous synthesis of organic molecules

SciAgents: Automating scientific discovery through multi-agent intelligent graph reasoning