Active learning and molecular dynamics simulations to find high melting temperature alloys

Farache, David E.; Verduzco, Juan Carlos; McClure, Zachary D.; Desai, Saaketh; Strachan, Alejandro

doi:10.1016/j.commatsci.2022.111386

Cited by 13 publications

(5 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It evaluates how much the prediction may change if a similar model had been trained on a different dataset. This measure was used for AL for example in [6] which propose a method to predict alloys melting points. It is expressed as :…”

Section: Proposalmentioning

confidence: 99%

Toward a Sawmill Digital Shadow Based on Coupled Simulation and Supervised Learning Models

Chabanet

Haouzi

Thomas

2023

Studies in Computational Intelligence

View full text Add to dashboard Cite

Digital Twins (DT) have been introduced as promising decision support tools in many different settings and to serve a variety of purposes. Many challenges are raised by their development, including an efficient usage of their computational resources to balance performance on precision, computational cost and speed. This study is, in particular, concerned with Digital Shadows (DS), a concept derived from DT, applied to sawmills sawing production units. A method to combine a computationally intensive sawmill simulation model with a machine learning model is proposed to predict set of lumbers sawed from logs. Numeric experiments are exposed, and the proposed method demonstrate improvements from 11% to 18% of the monitored couple regret from its baseline.

show abstract

Section: Proposalmentioning

confidence: 99%

Toward a Sawmill Digital Shadow Based on Coupled Simulation and Supervised Learning Models

Chabanet

Haouzi

Thomas

2023

Studies in Computational Intelligence

View full text Add to dashboard Cite

show abstract

“…Computational T m prediction has been demonstrated using various approaches (e.g., hysteresis, [17][18][19] two-phase coexistence, 20,21 interface pinning, 22,23 and other methods [24][25][26][27][28] ) and has been actively explored for decades. 24,25,[29][30][31][32][33][34][35][36] The two-phase coexistence (TPC) approach employing large supercells (>10 000 atoms) is considered the "gold standard" for predicting T m, as this approach relies on very few assumptions. 37 However, accurate computational predictions based on rst-principles methods, which are accurate and generally predictive, are prohibitively computationally expensive as ensuing calculations are "large" in two domains: the number of atoms in the simulation cell and the length of simulation time.…”

Section: Introductionmentioning

confidence: 99%

Predicting melting temperatures across the periodic table with machine learning atomistic potentials

Andolina,

Saidi

2024

Digital Discovery

View full text Add to dashboard Cite

show abstract

“…Since the uncertainty estimates are often associated with candidate materials outside the training space, improved model predictions may be found by iteratively sampling candidates that show either desired values of the target property or high uncertainty, and then retraining the surrogate model with these candidates included in the training data. Such an "active learning" workflow has been demonstrated to work well for the discovery of Ir-oxides 21 , transition metal complexes 22 , intermetallics 23 , transition metal dichalcogenides 24 , solid-state electrolytes 25 , and high melting temperature alloys 26 , among many others. In each of these cases, the active learning workflow is geared towards the optimization of a particular target property of interest along with improvement in model predictions.…”

Section: Introductionmentioning

confidence: 99%

Active Learning of Ternary Alloy Structures and Energies

Deshmukh

Wichrowski

Evangelou

et al. 2023

Preprint

View full text Add to dashboard Cite

High-throughput screening of catalysts using first-principles methods, such as density functional theory (DFT), has traditionally been limited by the large, complex, and multidimensional nature of the associated materials spaces. However, machine learning models with uncertainty quantification have recently emerged as attractive tools to accelerate the navigation of these spaces in a data-efficient manner, typically through active learning-based workflows. In this work, we combine such an active learning scheme with a dropout graph convolutional network (dGCN) as a surrogate model to explore the complex materials space of high-entropy alloys (HEAs). Specifically, we train the dGCN on the formation energies of disordered binary alloy structures in the Pd-Pt-Sn ternary alloy system and utilize the model to make and improve predictions on ternary structures. To do so, we perform reduced optimization over ensembles of ternary structures constructed based on two coordinate systems: (a) a physics-informed ternary composition space, and (b) data-driven coordinates discovered by the manifold learning scheme known as Diffusion Maps. Inspired by statistical mechanics, we derive and apply a dropout-informed acquisition function to select ensembles from which to sample additional structures. During each iteration of our active learning scheme, a representative number of crystals that minimize the acquisition function is selected, their energies are computed with DFT, and our dGCN model is retrained. We demonstrate that both of our reduced optimization techniques can be used to improve predictions of the formation free energy, the target property that determines HEA stability, in the ternary alloy space with a significantly reduced number of costly DFT calculations compared to a high-fidelity model. However, the manner in which these two disparate schemes converge to the target property differs: the physics-based scheme appears akin to a depth-first strategy, whereas the data-driven scheme appears more akin to a breadth-first approach. Both active learning schemes can be extended further to incorporate greater number of elements, surface structures, and adsorbate motifs.

show abstract

Active learning and molecular dynamics simulations to find high melting temperature alloys

Cited by 13 publications

References 58 publications

Toward a Sawmill Digital Shadow Based on Coupled Simulation and Supervised Learning Models

Toward a Sawmill Digital Shadow Based on Coupled Simulation and Supervised Learning Models

Predicting melting temperatures across the periodic table with machine learning atomistic potentials

Active Learning of Ternary Alloy Structures and Energies

Contact Info

Product

Resources

About