From Phase Prediction to Phase Design: A ReAct Agent Framework for High-Entropy Alloy Discovery

2026-03-10

AI Summary

This study addresses the inefficiency of inverse design for high-entropy alloys in high-dimensional compositional spaces by introducing, for the first time, a large language model (LLM) agent based on the ReAct framework. Integrated with an XGBoost surrogate model calibrated on 4,753 experimental data points, the agent autonomously proposes, validates, and iteratively optimizes alloy compositions targeting specific crystal phases such as FCC or BCC. By incorporating domain-specific prior knowledge, the approach maintains compositional diversity while closely adhering to the true phase manifold, clearly distinguishing between reproduced literature compositions and novel discoveries. Experimental results demonstrate a phase prediction accuracy of 94.66% (macro F1 = 0.896), with designed compositions lying 2.4–22.8 times closer to the experimental phase manifold than those from baseline methods like Bayesian optimization, and exhibiting significantly higher rediscovery rates.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

Discovering high-entropy alloy (HEA) compositions that reliably form a target crystal phase is a high-dimensional inverse design problem that conventional trial-and-error experimentation and forward-only machine learning models cannot efficiently solve. Here we present a ReAct (Reasoning + Acting) LLM agent that autonomously proposes, validates, and iteratively refines HEA compositions by querying a calibrated XGBoost surrogate trained on 4,753 experimental records across four phases (FCC, BCC, BCC+FCC, BCC+IM), achieving 94.66\% accuracy (F1 macro = 0.896). Against Bayesian optimisation (BO) and random search baselines, the full-prompt agent achieves descriptor-space rediscovery rates of 38\%, 18\%, and 38\% for FCC, BCC, and BCC+FCC (Mann--Whitney $p \leq 0.039$), with proposals lying 2.4--22.8$\times$ closer to the experimental phase manifold than random search. An ablation reveals that domain priors shift the agent from landmark-alloy recall toward compositionally diverse exploration -- an uninformed agent scores higher rediscovery by concentrating on literature-dense families, while the full-prompt agent explores underrepresented space (unique ratio 1.0 vs.\ 0.39 for BCC+FCC). These regimes represent distinct criteria: proximity to known literature versus genuine discovery. Spearman analysis confirms agent reasoning is statistically aligned with empirical phase distributions ($ρ= 0.736$, $p = 0.004$ for BCC). This work establishes LLM-guided agentic reasoning as a principled, transparent, and manifold-aware complement to gradient-free optimisation for inverse alloy design.

Problem

• high-entropy alloy
• phase design
• inverse design
• crystal phase
• composition discovery

Innovation

• ReAct agent
• high-entropy alloy
• inverse design
• LLM-guided reasoning
• phase prediction