Abstract:This study introduces a language transformer-based machine learning model to predict key mechanical properties of high-entropy alloys (HEAs), addressing the challenges due to their complex, multi-principal element compositions and limited experimental data. By pre-training the transformer on extensive synthetic materials data and fine-tuning it with specific HEA datasets, the model effectively captures intricate elemental interactions through self-attention mechanisms. This approach mitigates data scarcity issues via transfer learning, enhancing predictive accuracy for properties like elongation (%) and ultimate tensile strength (UTS) compared to traditional regression models such as Random Forests and Gaussian Processes. The model's interpretability is enhanced by visualizing attention weights, revealing significant elemental relationships that align with known metallurgical principles. This work demonstrates the potential of transformer models to accelerate materials discovery and optimization, enabling accurate property predictions, thereby advancing the field of materials informatics.

What problem does this paper attempt to address?

This paper attempts to solve several key problems in the prediction of mechanical properties of high - entropy alloys (HEAs): 1. **Challenges of complex multi - principal - element compositions**: High - entropy alloys are usually composed of five or more elements in near - equiatomic proportions, resulting in their complex compositional structures. Traditional design methods are difficult to effectively handle such multi - principal - element systems. 2. **Limited experimental data**: Due to experimental costs and time limitations, the experimental data of high - entropy alloys are relatively scarce, which makes the training based on traditional machine - learning models prone to overfitting and with poor generalization ability. 3. **Capturing complex interaction relationships between elements**: The long - range interactions between elements and the high - dimensional feature space in high - entropy alloys pose challenges to existing machine - learning models, which may not be able to fully capture these complex interaction relationships. To solve the above problems, the author proposes a Transformer - based language - model method to improve the prediction of mechanical properties of high - entropy alloys in the following ways: - **Combination of pre - training and fine - tuning**: First, pre - training is carried out on large - scale synthetic material data, and then fine - tuning is performed using a specific high - entropy alloy data set. This method takes advantage of transfer learning and alleviates the problem of data scarcity. - **Self - attention mechanism**: The self - attention mechanism in the Transformer model can effectively capture the complex interaction relationships between elements and improve the prediction accuracy. - **Enhanced interpretability**: By visualizing the attention weights, important element relationships are revealed, enhancing the interpretability of the model and making the results more in line with known metallurgical principles. Specifically, this research aims to predict the key mechanical properties of high - entropy alloys, such as elongation at break and ultimate tensile strength (UTS), and demonstrates the superior performance of the Transformer - based model in these tasks, especially in the case of limited data volume.

High Entropy Alloy property predictions using Transformer-based language model

Introducing pre-trained transformers for high entropy alloy informatics

Using Machine Learning and Feature Engineering to Characterize Limited Material Datasets of High-Entropy Alloys

Interpretable Machine Learning for High-Strength High-Entropy Alloy Design

Prediction of mechanical properties of non-equiatomic high-entropy alloy by atomistic simulation and machine learning

A neural network model for high entropy alloy design

Composition design of high-entropy alloys with deep sets learning

Prediction and design of high hardness high entropy alloy through machine learning

Predicting Elastic Properties of Refractory High-Entropy Alloys Via Machine-Learning Approach

Machine learning-assisted design of high-entropy alloys with superior mechanical properties

AlloyBERT: Alloy Property Prediction with Large Language Models

High-Entropy Alloys Properties Prediction Model by Using Artificial Neural Network Algorithm

Recent machine learning-driven investigations into high entropy alloys: A comprehensive review

Phase prediction and experimental realisation of a new high entropy alloy using machine learning

Machine learning for high-entropy alloys: Progress, challenges and opportunities

Predictive Modeling of High-Entropy Alloys and Amorphous Metallic Alloys Using Machine Learning

Predictive descriptors in machine learning and data-enabled explorations of high-entropy alloys

Prediction of the Composition and Hardness of High-Entropy Alloys by Machine Learning

Interpretable hardness prediction of high-entropy alloys through ensemble learning

Integrating machine learning with mechanistic models for predicting the yield strength of high entropy alloys

Machine Learning Assisted Design of High Entropy Alloys with Desired Property