Machine learning-unsupervised learning and supervised learning

1. Unsupervised learning
   has no labels , and the input category is not known before training. The training algorithm can classify it according to the common characteristics of the training set,
   such as k-means clustering algorithm
   
2. Supervised learning
   with a label , known in advance the training set should belong to a certain category
   such as prior knowledge of the training set is a cat or dog, you need to mark its kind, this is a very labor-consuming work.