The two terms "supervised learning" and "unsupervised learning" form a related pair of terms. That’s why both are explained here.
Supervised learning: any machine learning process in which the algorithm is somehow steered towards correct results using external resources. In practice, this means that well-curated training data is used for training.
Unsupervised learning: Any machine learning process that does not require prior knowledge to reach its conclusions. In other words: A procedure that is not guided but instead runs independently. The algorithm AlphaGo Zero was NOT fed with training data; the chess computer simply played millions of games against itself and derived strategies for chess from this. For your information: AlphaGo Zero (unsupervised learning) trounced the previous version AlphaGo (supervised learning).
