Semi-Supervised Learning - What is the difference between weak supervised and semi-supervised learning?

The main difference between weak supervised learning and semi-supervised learning is that the latter injects information into the system rather than propagating it, by making use of pre-existing labels to create new ones.