From Data to Discovery: Machine Learning–Enabled Intelligent Characterization of Two‐Dimensional Materials
With the rapid development of two‐dimensional (2D) materials, characterization techniques are progressively achieving atomic‐scale accuracy and intelligent automation. Conventional methods usually rely on manual and experience‐based analysis, which is insufficient to meet the demands of diverse material systems, massive multimodal datasets, and reproducible quantitative analysis. This paper summarizes recent progress in the application of machine learning (ML) algorithms to 2D material characterization, including optical microscopy, photoluminescence (PL), Raman spectroscopy, scanning/transmission electron microscopy (STEM/TEM), and scanning probe microscopy (SPM). The main advances include automatic identification and quantitative analysis of layered structures, strain, and defects, as well as the correlation between these structural features and complex spectral data. In addition, active learning strategies can guide the experimental workflow and enable intelligent operation of characterization instruments. Despite these promising advances, several challenges remain, including the construction of large‐scale high‐quality datasets, limited model interpretability, and insufficient cross‐platform generalization capability.