An official website of the United States government
Abstract In real-world materials research, machine learning (ML) models are usually expected to predict and discover novel exceptional materials that deviate from the known materials. It is thus a pressing question to provide an objective evaluation of ML model performances in property prediction of out-of-distribution (OOD) materials that are different from the training set. Traditional performance evaluation of materials property prediction models through the random splitting of the dataset frequently results in artificially high-performance assessments due to the inherent redundancy of typical material datasets. Here we present a comprehensive benchmark study of structure-based graph neural networks (GNNs) for extrapolative OOD materials property prediction. We formulate five different categories of OOD ML problems for three benchmark datasets from the MatBench study. Our extensive experiments show that current state-of-the-art GNN algorithms significantly underperform for the OOD property prediction tasks on average compared to their baselines in the MatBench study, demonstrating a crucial generalization gap in realistic material prediction tasks. We further examine the latent physical spaces of these GNN models and identify the sources of CGCNN, ALIGNN, and DeeperGATGNN’s significantly more robust OOD performance than those of the current best models in the MatBench study (coGN and coNGN) as a case study for the perovskites dataset, and provide insights to improve their performance.
more »
« less
Advancing flame retardant prediction: A self-enforcing machine learning approach for small datasets
Improving the fireproof performance of polymers is crucial for ensuring human safety and enabling future space colonization. However, the complexity of the mechanisms for flame retardant and the need for customized material design pose significant challenges. To address these issues, we propose a machine learning (ML) framework based on substructure fingerprinting and self-enforcing deep neural networks (SDNN) to predict the fireproof performance of flame-retardant epoxy resins. Our model is based on a comprehensive understanding of the physical mechanisms of materials and can predict fireproof performance and eliminate the needs for properties descriptors, making it more convenient than previous ML models. With a dataset of only 163 samples, our SDNN models show an average prediction error of 3% for the limited oxygen index (LOI). They also provide satisfactory predictions for the peak of heat release rate PHR and total heat release (THR), with coefficient of determination (R2) values of 0.87 and 0.85, respectively, and average prediction errors less than 17%. Our model outperforms the support vector model SVM for all three indices, making it a state-of-the-art study in the field of flame retardancy. We believe that our framework will be a valuable tool for the design and virtual screening of flame retardants and will contribute to the development of safer and more efficient polymer materials.
more »
« less
- Award ID(s):
- 1946231
- PAR ID:
- 10496338
- Publisher / Repository:
- https://pubs.aip.org/aip/apl/article-abstract/122/25/251902/2897699/Advancing-flame-retardant-prediction-A-self?redirectedFrom=PDF
- Date Published:
- Journal Name:
- Applied Physics Letters
- Volume:
- 122
- Issue:
- 25
- ISSN:
- 0003-6951
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract One compelling vision of the future of materials discovery and design involves the use of machine learning (ML) models to predict materials properties and then rapidly find materials tailored for specific applications. However, realizing this vision requires both providing detailed uncertainty quantification (model prediction errors and domain of applicability) and making models readily usable. At present, it is common practice in the community to assess ML model performance only in terms of prediction accuracy (e.g. mean absolute error), while neglecting detailed uncertainty quantification and robust model accessibility and usability. Here, we demonstrate a practical method for realizing both uncertainty and accessibility features with a large set of models. We develop random forest ML models for 33 materials properties spanning an array of data sources (computational and experimental) and property types (electrical, mechanical, thermodynamic, etc). All models have calibrated ensemble error bars to quantify prediction uncertainty and domain of applicability guidance enabled by kernel-density-estimate-based feature distance measures. All data and models are publicly hosted on the Garden-AI infrastructure, which provides an easy-to-use, persistent interface for model dissemination that permits models to be invoked with only a few lines of Python code. We demonstrate the power of this approach by using our models to conduct a fully ML-based materials discovery exercise to search for new stable, highly active perovskite oxide catalyst materials.more » « less
-
Abstract Fire retardant coatings have been proven effective at reducing the heat release rate (HRR) of structural materials during burning; yet effective methods for increasing the ignition temperature and delay time prior to burning are rarely reported. Herein, a strong, fire‐resistant wood structural material is developed by combining a densification treatment with an anisotropic thermally conductive flame‐retardant coating of hexagonal boron nitride (h‐BN) nanosheets to produce BN‐densified wood. The thermal management properties created by the BN coating provide fast, in‐plane thermal diffusion, slowing the conduction of heat through the densified wood, which improves the material's ignition properties. Compared with densified wood without the BN coating, a 41 °C enhancement in ignition temperature (Tig), a twofold increase in ignition delay time (tig), and a 25% decrease in the maximum HRR of BN‐densified wood can be achieved. As a proof of concept for scalability, the pieces of the BN‐densified wood are fabricated with a length larger than 25 cm, width greater than 15 cm, and thickness more than 7 mm. The improved thermal management, fire resistance, mechanical strength, and scalable production of BN‐densified wood position it as a promising structural material for safe and energy‐efficient buildings.more » « less
-
Flame-retardant and thermal management structures have attracted great attention due to the requirement of high-temperature exposure in industrial, aerospace, and thermal power fields, but the development of protective fire-retardant structures with complex shapes to fit arbitrary surfaces is still challenging. Herein, we reported a rotation-blade casting-assisted 3D printing process to fabricate nacre-inspired structures with exceptional mechanical and flame-retardant properties, and the related fundamental mechanisms are studied. 3-(Trimethoxysilyl)propyl methacrylate (TMSPMA) modified boron nitride nanoplatelets (BNs) were aligned by rotation-blade casting during the 3D printing process to build the “brick and mortar” architecture. The 3D printed structures are more lightweight, while having higher fracture toughness than the natural nacre, which is attributed to the crack deflection, aligned BN (a-BNs) bridging, and pull-outs reinforced structures by the covalent bonding between TMSPMA grafted a-BNs and polymer matrix. Thermal conductivity is enhanced by 25.5 times compared with pure polymer and 5.8 times of anisotropy due to the interconnection of a-BNs. 3D printed heat-exchange structures with vertically aligned BNs in complex shapes were demonstrated for efficient thermal control of high-power light-emitting diodes. 3D printed helmet and armor with a-BNs show exceptional mechanical and fire-retardant properties, demonstrating integrated mechanical and thermal protection.more » « less
-
Abstract Modern machine learning (ML) and deep learning (DL) techniques using high-dimensional data representations have helped accelerate the materials discovery process by efficiently detecting hidden patterns in existing datasets and linking input representations to output properties for a better understanding of the scientific phenomenon. While a deep neural network comprised of fully connected layers has been widely used for materials property prediction, simply creating a deeper model with a large number of layers often faces with vanishing gradient problem, causing a degradation in the performance, thereby limiting usage. In this paper, we study and propose architectural principles to address the question of improving the performance of model training and inference under fixed parametric constraints. Here, we present a general deep-learning framework based on branched residual learning (BRNet) with fully connected layers that can work with any numerical vector-based representation as input to build accurate models to predict materials properties. We perform model training for materials properties using numerical vectors representing different composition-based attributes of the respective materials and compare the performance of the proposed models against traditional ML and existing DL architectures. We find that the proposed models are significantly more accurate than the ML/DL models for all data sizes by using different composition-based attributes as input. Further, branched learning requires fewer parameters and results in faster model training due to better convergence during the training phase than existing neural networks, thereby efficiently building accurate models for predicting materials properties.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1063/5.0152195
