Semi-Supervised Learning

Not to be confused with Self-Supervised Learning.

I’ve never heard this being used much.

  • Definition: Semi-supervised learning sits between supervised and unsupervised learning. It uses a small amount of labeled data alongside a larger pool of unlabeled data.
  • Approach: The labeled data provides initial guidance on the task, and the model then extends this understanding to the unlabeled data, often by understanding the underlying structure or distribution of the data.
  • Applications: It’s beneficial when labeled data is limited or costly to obtain but unlabeled data is abundant. Common applications include text classification, image recognition, and more complex tasks like medical image analysis.