[Paper Review] Agentic AI and Machine Learning for Accelerated Materials Discovery and Applications

Jihua Chen, Panagiotis Christakopoulos+3 more|arXiv (Cornell University)|Jan 13, 2026

Machine Learning in Materials Science0 citations

TL;DR

The paper surveys agentic AI and ML for faster materials discovery, highlighting core concepts, efficiency needs, and examples in flow chemistry, biosensors, and batteries with a focus on large language models.

ABSTRACT

Artificial Intelligence (AI), especially AI agents, is increasingly being applied to chemistry, healthcare, and manufacturing to enhance productivity. In this review, we discuss the progress of AI and agentic AI in areas related to, and beyond polymer materials and discovery chemistry. More specifically, the focus is on the need for efficient discovery, core concepts, and large language models. Consequently, applications are showcased in scenarios such as (1) flow chemistry, (2) biosensors, and (3) batteries.

Motivation & Objective

• Assess the role and progress of AI agents in chemistry, healthcare, and manufacturing with emphasis on materials discovery.
• Identify challenges and requirements for efficient discovery in materials science.
• Discuss core concepts of agentic AI and the role of large language models in this domain.
• Showcase practical applications in flow chemistry, biosensors, and batteries.

Proposed method

• Review existing literature on agentic AI and ML in materials science and related fields.
• Theoretically discuss efficient discovery workflows and decision-making processes in AI agents.
• Highlight the relevance and capabilities of large language models for materials discovery tasks.

Experimental results

Research questions

• RQ1 What are the current advances and limitations of agentic AI in accelerated materials discovery?
• RQ2 How can large language models and AI agents be leveraged to improve discovery workflows in materials science?
• RQ3 What practical applications (e.g., flow chemistry, biosensors, batteries) illustrate the potential of agentic AI in this domain?

Key findings

• AI agents and ML are increasingly applied to materials discovery and related areas to boost productivity.
• Efficient discovery workflows and core concepts underpin successful agentic AI implementations.
• Large language models are a focal point for enabling reasoning, planning, and data interaction in materials discovery.
• Applications are demonstrated in scenarios such as flow chemistry, biosensors, and batteries.