Hi,

In supervised learning the label (class) of the instances is known (training data are labeled). So, each example or instance has a specifical label

Unsupervised learning, the label of instances is not known (training data are not labeled). In thise case, we must first determine the number of classes.

The network learning in a controlled way, namely on the basis selected by the "trainer", as well as formulated by the assessments.

Unsupervised learning without a teacher and in direct interaction with the environment. This kind of networks are implemented with some algorithms of the network structure. They are quite general and non-targeted to a specific purpose, but inspired by some general laws of adaptation (as living organisms). The lack of human trainer over the network during the arriving of stimuli, the results of these algorithms or changes in the structure of the network, are unpredictable.

Supervised learning is used for classification problems where you have the data and the label for the data to which class it belongs to. unsupervised learning is meant for clustering of data, estimation techniques etc where probabilistic reasoning is used to make a decision.

Supervised machine learning is a inferring a function from label training data, but unsupervised machine learning is that trying to find hidden structure in unlabeled data.

see link

https://en.wikipedia.org/wiki/Supervised_learning

Supervised ML algorithms predict a solution for a regression or classification data problem by using information about the same problem (training dataset); however, unsupervised ML algorithms predict or discover hidden patterns or clusters in data without any training datasets.