Learning Molecular Mixture Property Using Chemistry-Aware Graph Neural Network DOI Creative Commons
Hengrui Zhang, Tianxing Lai, Jie Chen

et al.

PRX Energy, Journal Year: 2024, Volume and Issue: 3(2)

Published: June 12, 2024

Recent advances in machine learning (ML) are expediting materials discovery and design. One significant challenge facing ML for is the expansive combinatorial space of potential formed by diverse constituents their flexible configurations. This complexity particularly evident molecular mixtures, a frequently explored materials, such as battery electrolytes. Owing to complex structures molecules sequence-independent nature conventional methods have difficulties modeling systems. Here, we present MolSets, specialized model overcome difficulties. Representing individual graphs mixture set, MolSets leverages graph neural network deep sets architecture extract information at level aggregate it level, thus addressing local while retaining global flexibility. We demonstrate efficacy predicting conductivity lithium electrolytes highlight its benefits virtual screening chemical space. Published American Physical Society 2024

Language: Английский

Integrating artificial intelligence in energy transition: A comprehensive review DOI Creative Commons
Qiang Wang,

Yuanfan Li,

Rongrong Li

et al.

Energy Strategy Reviews, Journal Year: 2025, Volume and Issue: 57, P. 101600 - 101600

Published: Jan. 1, 2025

Language: Английский

Citations

24
Rise of machine learning potentials in heterogeneous catalysis: Developments, applications, and prospects DOI
Seokhyun Choung,

Wongyu Park,

Jinuk Moon

et al.

Chemical Engineering Journal, Journal Year: 2024, Volume and Issue: 494, P. 152757 - 152757

Published: June 2, 2024

Language: Английский

Citations

15
Leveraging language representation for materials exploration and discovery DOI Creative Commons
Jiaxing Qu, Yuxuan Richard Xie, Kamil Ciesielski

et al.

npj Computational Materials, Journal Year: 2024, Volume and Issue: 10(1)

Published: March 21, 2024

Abstract Data-driven approaches to materials exploration and discovery are building momentum due emerging advances in machine learning. However, parsimonious representations of crystals for navigating the vast search space remain limited. To address this limitation, we introduce a framework that utilizes natural language embeddings from models as compositional structural features. The contextual knowledge encoded these conveys information about material properties structures, enabling both similarity analysis recall relevant candidates based on query multi-task learning share across related properties. Applying thermoelectrics, demonstrate diversified recommendations prototype crystal structures identify under-studied spaces. Validation through first-principles calculations experiments confirms potential recommended high-performance thermoelectrics. Language-based frameworks offer versatile adaptable embedding effective discovery, applicable diverse systems.

Language: Английский

Citations

13
Towards atom-level understanding of metal oxide catalysts for the oxygen evolution reaction with machine learning DOI Creative Commons

Jaclyn R. Lunger,

Jessica Karaguesian,

Hoje Chun

et al.

npj Computational Materials, Journal Year: 2024, Volume and Issue: 10(1)

Published: April 22, 2024

Abstract Green hydrogen production is crucial for a sustainable future, but current catalysts the oxygen evolution reaction (OER) suffer from slow kinetics, despite many efforts to produce optimal designs, particularly through calculation of descriptors activity. In this study, we develop dataset density functional theory calculations bulk and surface perovskite oxides, adsorption energies OER intermediates, which includes compositions up quaternary facets (555). We demonstrate that per-site properties oxides such as Bader charge or band center can be tuned element substitution faceting, machine learning model accurately predicts these directly local chemical environment. leverage identify promising perovskites with high theoretical The identified design principles materials provide roadmap closing gap between artificial biological enzymes photosystem II.

Language: Английский

Citations

13
Accelerating the prediction of inorganic surfaces with machine learning interatomic potentials DOI
Kyle Noordhoek, Christopher J. Bartel

Nanoscale, Journal Year: 2024, Volume and Issue: 16(13), P. 6365 - 6382

Published: Jan. 1, 2024

This minireview summarizes recent applications of machine learning interatomic potentials for predicting the stability and structures solid-state surfaces.

Language: Английский

Citations

10
Machine-learning-accelerated simulations to enable automatic surface reconstruction DOI
Xiaochen Du,

James K. Damewood,

Jaclyn R. Lunger

et al.

Nature Computational Science, Journal Year: 2023, Volume and Issue: 3(12), P. 1034 - 1044

Published: Dec. 7, 2023

Language: Английский

Citations

22
Methods and applications of machine learning in computational design of optoelectronic semiconductors DOI Open Access
Xiaoyu Yang, Kun Zhou, Xin He

et al.

Science China Materials, Journal Year: 2024, Volume and Issue: 67(4), P. 1042 - 1081

Published: March 19, 2024

Language: Английский

Citations

8
Predictive Modeling of Abrasive Wear in In-Situ TiC Reinforced ZA37 Alloy: A Machine Learning Approach DOI
Khursheed Ahmad Sheikh, Mohammad Mohsin Khan

Tribology International, Journal Year: 2024, Volume and Issue: unknown, P. 110291 - 110291

Published: Sept. 1, 2024

Language: Английский

Citations

6
Machine learning-accelerated discovery of heat-resistant polysulfates for electrostatic energy storage DOI
He Li, Hongbo Zheng, Tianle Yue

et al.

Nature Energy, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 5, 2024

Language: Английский

Citations

6
When Metal Nanoclusters Meet Smart Synthesis DOI
Zhucheng Yang, Anye Shi, Ruixuan Zhang

et al.

ACS Nano, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 24, 2024

Atomically precise metal nanoclusters (MNCs) represent a fascinating class of ultrasmall nanoparticles with molecule-like properties, bridging conventional metal-ligand complexes and nanocrystals. Despite their potential for various applications, synthesis challenges such as understanding varied synthetic parameters property-driven persist, hindering full exploitation wider application. Incorporating smart methodologies, including closed-loop framework automation, data interpretation, feedback from AI, offers promising solutions to address these challenges. In this perspective, we summarize the that has been demonstrated in nanomaterials explore research frontiers MNCs. Moreover, perspectives on inherent opportunities MNCs are discussed, aiming provide insights directions future advancements emerging field AI Science, while integration deep learning algorithms stands substantially enrich by offering enhanced predictive capabilities, optimization strategies, control mechanisms, thereby extending MNC synthesis.

Language: Английский

Citations

5