Transforming Molecular Synthesis With Large Language Models.

Area of Science:

• Chemistry
• Artificial Intelligence
• Computational Chemistry

Background:

• Traditional data-driven methods in molecular synthesis are being augmented by advanced AI.
• Large language models (LLMs) offer unique capabilities in knowledge absorption and logical inference.

Purpose of the Study:

• To review the transformative impact of LLMs on molecular synthesis.
• To outline the key roles LLMs play in automating synthesis workflows.
• To provide a roadmap for leveraging LLMs in chemical research.

Main Methods:

• The review systematically analyzes the application of LLMs across four primary functions.
• LLMs are examined as knowledge bases, data structuring tools, predictive models, and intelligent agents.

Main Results:

• LLMs serve as interactive knowledge bases for expert consultation.
• LLMs effectively structure unstructured reaction data from scientific literature.
• LLMs demonstrate versatility as predictors for molecular properties and reaction outcomes.
• LLMs function as central components in intelligent agent systems for autonomous synthesis.

Conclusions:

• LLMs represent a paradigm shift in molecular synthesis, moving beyond traditional approaches.
• The integration of LLMs accelerates the development of automated and intelligent synthesis planning and execution.
• This review provides a comprehensive guide for chemists to utilize LLM capabilities.