From bulk effective mass to 2D carrier mobility accurate prediction via adversarial transfer learning

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: June 25, 2024

Abstract Data scarcity is one of the critical bottlenecks to utilizing machine learning in material discovery. Transfer can use existing big data assist property prediction on small sets, but premise that there must be a strong correlation between large and sets. To extend its applicability scenarios with different properties materials, here we develop hybrid framework combining adversarial transfer expert knowledge, which enables direct carrier mobility two-dimensional (2D) materials using knowledge learned from bulk effective mass. Specifically, training ensures only common 2D extracted while incorporated further improve accuracy generalizability. Successfully, mobilities are predicted over 90% crystal structure, 21 semiconductors far exceeding silicon suitable bandgap successfully screened out. This work simultaneous cross-property cross-material scenarios, providing an tool predict intricate limited data.

Language: Английский

---

Applications of machine learning in surfaces and interfaces

Chemical Physics Reviews, Journal Year: 2025, Volume and Issue: 6(1)

Published: March 1, 2025

Surfaces and interfaces play key roles in chemical material science. Understanding physical processes at complex surfaces is a challenging task. Machine learning provides powerful tool to help analyze accelerate simulations. This comprehensive review affords an overview of the applications machine study systems materials. We categorize into following broad categories: solid–solid interface, solid–liquid liquid–liquid surface solid, liquid, three-phase interfaces. High-throughput screening, combined first-principles calculations, force field accelerated molecular dynamics simulations are used rational design such as all-solid-state batteries, solar cells, heterogeneous catalysis. detailed information on for

Language: Английский

---

Engineering interfacial charge transfer through modulation doping for 2D electronics

Physical Review Materials, Journal Year: 2025, Volume and Issue: 9(2)

Published: Feb. 18, 2025

Language: Английский

---

Machine Learning-Assisted Dielectric Screening of Bismuth/Antimony-Based Compounds for Promising Optoelectronic Semiconductors

The Journal of Physical Chemistry C, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 27, 2025

Language: Английский

---

Point Defects Analysis and Surface Property Prediction by Machine Learning

Materia Japan, Journal Year: 2025, Volume and Issue: 64(3), P. 184 - 189

Published: Feb. 28, 2025

Language: Английский

---

Surface reconstruction and band alignment of non-metallic spinel oxides from first principles

Acta Materialia, Journal Year: 2025, Volume and Issue: unknown, P. 121034 - 121034

Published: April 1, 2025

Language: Английский

---

Bayesian Optimization with Gaussian Processes Assisted by Deep Learning for Material Designs

The Journal of Physical Chemistry Letters, Journal Year: 2025, Volume and Issue: unknown, P. 5244 - 5251

Published: May 18, 2025

Machine learning (ML) approaches have become ubiquitous in the search for new materials recent years. Bayesian optimization (BO) based on Gaussian processes (GPs) has a widely recognized approach material exploration. However, feature engineering critical impacts efficiency of GP-based BO, because GPs cannot automatically generate descriptors. To address this limitation, study applies deep kernel (DKL), which combines neural network with GP, to BO. The DKL model was comparable or significantly better than that standard GP data set 922 oxide sets, covering band gaps, ionic dielectric constants, and effective masses electrons, as well experimental gaps 610 hybrid organic-inorganic perovskite alloys. When searching alloy highest Curie temperature among 4560 alloys, outperformed strongly correlated descriptor could be directly utilized. Additionally, supports transfer learning, further enhances its efficiency. Thus, we believe BO paves way exploring diverse spaces more effectively GPs.

Language: Английский

---

Theoretical and data-driven approaches to semiconductors and dielectrics: from prediction to experiment

Science and Technology of Advanced Materials, Journal Year: 2024, Volume and Issue: 25(1)

Published: Nov. 4, 2024

Computational approaches using theoretical calculations and data scientific methods have become increasingly important in materials science technology, with the development of relevant methodologies algorithms, availability large data, enhancement computer performance. As reviewed herein, we developed computational for design prediction inorganic a particular focus on exploration semiconductors dielectrics. High-throughput first-principles are used to systematically accurately predict local atomic electronic structures polarons, point defects, surfaces, interfaces, as well bulk fundamental properties. Machine learning techniques utilized efficiently various material properties, construct phase diagrams, search satisfying target These elucidated mechanisms behind functionalities explored promising combination synthesis, characterization, device fabrication. Examples include ternary nitride potential optoelectronic photovoltaic applications, phosphide optimization heterointerfaces toward improvement phosphide-based cells, discovery ferroelectricity layered perovskite oxides understanding its origin, all which demonstrate effectiveness our computer-aided research.

Language: Английский

---

A surface emphasized multi-task learning framework for surface property predictions: A case study of magnesium intermetallics

Journal of Magnesium and Alloys, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 1, 2024

Language: Английский

---