Semi-supervised CAPP Transformer Learning via Pseudo-labeling
Experiments on small-scale datasets with simulated ground truth across the full data distribution show consistent accuracy gains over baselines, demonstrating the method's effectiveness in data-scarce manufacturing environments.