Metalearning‐Driven Inverse Optimization for Precision Microstructure Fabrication in Digital Light Processing Three‐Dimensional Printing

Due to the nonlinear and coupled interactions between process parameters and environmental conditions, achieving dimensional precision in microstructure fabrication through digital light processing (DLP) printing requires a lot of trial and error. To develop user‐friendly DLP printing technology, it is necessary to optimize the DLP process using a data‐driven strategy. In this study, a metalearning‐enhanced neural network framework for the inverse design and optimization of cure depth and width in microscale DLP printing is proposed. A comprehensive experimental dataset is constructed using a bridge‐structure design, incorporating key variables such as UV‐exposure time, mask dimensions, bridge length, temperature, and humidity. Analysis of the dataset and model predictions reveals a consistent trend of overcured thickness and underfabricated width, primarily influenced by exposure time and mask width. To resolve these opposing deviations, an overmasking strategy combined with exposure‐time tuning is introduced, significantly improving fabrication accuracy. Validation using multilayer microstructures and scaffold arrays demonstrates that the dimensional error in cure depth is significantly reduced to below 5% compared to the previous level. The results highlight the effectiveness of machine learning‐guided process control in improving the precision, reproducibility, and scalability of DLP‐based microstructure fabrication, particularly for applications in tissue engineering, biomedical scaffolds, and high‐resolution additive manufacturing.