Abstract:We present an algorithm for computing melting points by autonomously learning from coexistence simulations in the NPT ensemble. Given the interatomic interaction model, the method makes decisions regarding the number of atoms and temperature at which to conduct simulations, and based on the collected data predicts the melting point along with the uncertainty, which can be systematically improved with more data. We demonstrate how incorporating physical models of the solid-liquid coexistence evolution enhances the algorithm's accuracy and enables optimal decision-making to effectively reduce predictive uncertainty. To validate our approach, we compare the results of 20 melting point calculations from the literature to the results of our calculations, all conducted with same interatomic potentials. Remarkably, we observe significant deviations in about one-third of the cases, underscoring the need for accurate and reliable algorithms for materials property calculations.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to develop an algorithm that can independently learn from co - existence simulations and predict the melting points of materials. Specifically, based on the given inter - atomic interaction model, this method independently determines the number of atoms and the temperature for the simulation, and predicts the melting point and its uncertainty according to the collected data. By introducing a physical model to describe the mechanism of solid - liquid co - existence evolution, this algorithm not only improves the prediction accuracy, but also can effectively reduce the prediction uncertainty through optimized decision - making. The key innovation points of the paper are as follows: 1. **Autonomous learning**: The algorithm can independently determine under which conditions to conduct molecular dynamics (MD) simulations in order to collect the most effective data. 2. **Integration of physical models**: By combining physical models, especially the Fokker - Planck equation for solid - liquid interface motion, the accuracy and reliability of the algorithm are improved. 3. **Uncertainty quantification**: Using the Bayesian regression method, it not only predicts the melting point, but also provides the confidence interval of the prediction, which is crucial for the calculation of material properties. 4. **Systematic improvement**: As more data are collected, the prediction uncertainty can be systematically reduced. To verify the effectiveness of this method, the author applied this method to the melting point calculations of 20 materials and compared the results with those in the literature. The results show that in about one - third of the cases, the prediction results of the new method are significantly different from those of the existing methods, which highlights the importance of developing accurate and reliable algorithms for calculating material properties.

Accurate melting point prediction through autonomous physics-informed learning

High-accuracy thermodynamic properties to the melting point from ab initio calculations aided by machine-learning potentials

Melting temperature prediction via first principles and deep learning

Discovering Melting Temperature Prediction Models of Inorganic Solids by Combining Supervised and Unsupervised Learning

A Generic and Automated Methodology to Simulate Melting Point

Toward Fully in Silico Melting Point Prediction Using Molecular Simulations.

Phthalonitrile Melting Point Prediction Enabled by Multi-Fidelity Learning

MeltNet: Predicting alloy melting temperature by machine learning

P2MAT: A machine learning (ML) driven software for Property Prediction of MATerial

Active Learning Prediction and Experimental Confirmation of Atomic Structure and Thermophysical Properties for Liquid Hf76 W24 Refractory Alloy

Atomistic simulation assisted error-inclusive Bayesian machine learning for probabilistically unraveling the mechanical properties of solidified metals

Melting temperature prediction using a graph neural network model: From ancient minerals to new materials

Can we accurately calculate viscosity in multicomponent metallic melts?

Physics-Informed Machine Learning for Accurate Prediction of Temperature and Melt Pool Dimension in Metal Additive Manufacturing

Tutorial: Deep Learning Prediction of Thermophysical Properties for Liquid Multicomponent Alloys

Efficient and accurate simulation of vitrification in multi-component metallic liquids with neural-network potentials

Rapid acquisition of liquid thermophysical properties from pure metals to quaternary alloys by proposing a machine learning strategy

Deciphering melting behaviors of energetic compounds using interpretable Machine learning for melt-castable applications

Accelerating ab initio melting property calculations with machine learning: Application to the high entropy alloy TaVCrW

A comprehensive investigation on the accuracy and efficiency of methods for melting temperature calculation using molecular dynamics simulations

Accelerating material melting temperature predictions by implementing machine learning potentials in the SLUSCHI package