Taking a Better Look at Data to Predict Exceptional Materials
© The Physical Society of Japan
This article is on
J. Phys. Soc. Jpn.
89,
064802
(2020)
.
Instead of regression, the evaluation values to measure the degree of separation of the distributions of the target variable were studied and applied to the magneto crystallineanisotropy of Co/Fe films.
J. Phys. Soc. Jpn.
89,
064802
(2020)
.
Share this topic
Fields
Related Articles
-
Exploring the Vibrant Interplay of Machine Learning and Physics
Cross-disciplinary physics and related areas of science and technology
Electron states in condensed matter
Elementary particles, fields, and strings
Mathematical methods, classical and quantum physics, relativity, gravitation, numerical simulation, computational modeling
Statistical physics and thermodynamics
Superconductivity
2025-3-13
This Journal of the Physical Society of Japan Special Topics edition explores how physics and machine learning complement each other and can solve unresolved problems in physics.
-
Understanding Pressure-Induced Superconductivity in CrAs and MnP
Magnetic properties in condensed matter
2025-3-10
This study reviews existing research on the pressure-induced variation of magnetic properties of transition metal mono-pnictides like CrAS, MnP, and others, aiming to understand the unconventional superconductivity observed in CrAs and MnP.
-
Symmetry and AI: Building the Future of Physics Simulations
Magnetic properties in condensed matter
Measurement, instrumentation, and techniques
2025-2-18
Generative artificial intelligence (AI) has gained considerable attention in scientific fields. By embedding physical symmetry into AI before training, we created a faster and lighter model. Scaling improves the accuracy and unlocks the potential of physics research and applications.
-
Exploring Materials without Data Exposure: A Bayesian Optimizer using Secure Computation
Cross-disciplinary physics and related areas of science and technology
Measurement, instrumentation, and techniques
2025-2-6
Secure computation allows the manipulation of material data without exposing them, thereby offering an alternative to traditional open/closed data management. We recently reported the development of an application that performs Bayesian optimization using secure computation.
-
Triangular Lattice Magnet GdGa2: Spin Cycloids and Skyrmions
Cross-disciplinary physics and related areas of science and technology
Electronic transport in condensed matter
Magnetic properties in condensed matter
2025-2-3
Careful measurements were conducted on the hexagonal magnet GdGa2 to reveal the experimental signatures of ultrasmall spin cycloids and of a potential Néel-type skyrmion lattice phase induced by a magnetic field.