Learning methods

• Weak supervision is supervision with noisy labels. For example, bootstrapping, where the bootstrapping procedure may mislabel some examples.
• Distant supervision refers to training signals that do not directly label the examples; for example, learning semantic parsers from question-and-answer datasets.
• Semi-supervised learning is when you have a dataset that is partially labeled and partially unlabeled.
• Full-supervised learning is when you have ground truth labels for each datapoint.

from Tudor Achim