Predicting magnetism with first-principles AI

2026-02-09

AI Summary

This work addresses the formidable challenge of modeling magnetic materials, which arises from the delicate competition between kinetic energy and Coulomb interactions—often beyond the reach of conventional electronic structure methods. The authors propose a neural-network-based variational Monte Carlo approach that directly solves the many-electron Schrödinger equation, employing highly expressive variational wave functions to accurately capture magnetic states in strongly correlated systems. Notably, their method simultaneously describes both ferromagnetic and antiferromagnetic order within a single spin sector (Sz = 0), circumventing the need for multi-sector calculations and substantially reducing computational cost. The approach successfully predicts itinerant ferromagnetism in WSe₂/WS₂ moiré semiconductors and an antiferromagnetic insulating state in twisted Γ-valley homobilayers, demonstrating its accuracy and efficiency, and establishing a new paradigm for the design of magnetic materials.

Technology Category

• Machine Learning: Matrix & Tensor Methods
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Multiagent Systems: Adversarial Agents

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Computational discovery of magnetic materials remains challenging because magnetism arises from the competition between kinetic energy and Coulomb interaction that is often beyond the reach of standard electronic-structure methods. Here we tackle this challenge by directly solving the many-electron Schr\"odinger equation with neural-network variational Monte Carlo, which provides a highly expressive variational wavefunction for strongly correlated systems. Applying this technique to transition metal dichalcogenide moir\'e semicondutors, we predict itinerant ferromagnetism in WSe$_2$/WS$_2$ and an antiferromagnetic insulator in twisted $\Gamma$-valley homobilayer, using the same neural network without any physics input beyond the microscopic Hamiltonian. Crucially, both types of magnetic states are obtained from a single calculation within the $S_z=0$ sector, removing the need to compute and compare multiple $S_z$ sectors. This significantly reduces computational cost and paves the way for faster and more reliable magnetic material design.

Problem

• magnetism
• first-principles
• strongly correlated systems
• magnetic materials
• electronic structure

Innovation

• neural-network variational Monte Carlo
• many-electron Schrödinger equation
• strongly correlated systems
• magnetic materials discovery
• Sz=0 sector prediction