This review explores the integration of artificial intelligence (AI) with condensed matter physics, specifically focusing on the prediction of magnetic higher-order topological corner states. We examine the theoretical foundations of higher-order topological insulators (HOTIs), the unique challenges posed by magnetic systems, and the application of equivariant graph neural networks in high-throughput screening. The article discusses active learning strategies for navigating vast chemical configuration spaces and addresses current limitations regarding data scarcity and strong electronic correlations.
Liang Chengtian (2026) studied this question.