Semi-Supervised Learning

This training method uses a small amount of labeled data combined with a large amount of unlabeled data. It's particularly useful when obtaining labeled data is expensive or time-consuming, allowing models to learn from both structured and unstructured information.