Machine Learning Types #supervisedlearning #unsupervisedlearning #ai #deeplearning #machinelearning

Ajouté : 2026-05-05

[00:00:01]Hi, good evening everyone. Welcome to another episode of AIQ. And today we are going to talk about the different type of machine learning. The first type of machine learning that we have that is basically called as [snorts] supervised machine learning.

[00:00:15]Supervised machine learning is where we go ahead and already tell or give supervision to the machine. So for example, we know that there are two kind of mails. One is a spam mail and one is a non-spam mail. Now I already train the machine with the label of the emails. So my already but I think the machine going already telling the machine that the first mail is a spam mail, the second mail is a non-spam mail, the third mail is a spam mail. I'm already telling the machine which mail is a spam and which mail is non-spam. And after giving it enough training for new mails, I'm asking the machine to go ahead and categorize it in spam versus non-spam. So since the supervision is already given, it is called as supervised machine learning.

[00:00:58]The second one that we have that is called as unsupervised machine learning.

[00:01:02]Unsupervised machine learning as the name suggests, there is no supervision given to the machine. So for example, I don't know what kind of consumers are there in the market for my product. So what I go ahead and I do is I give data to the machine and I ask the machine to go ahead and create market segments based on the hidden patterns that the machine goes ahead and observes. So if the machine sees that there is some pattern in one kind of consumers, it labels it uh consumer type A. Then again, it sees the other type of consumer and it sees some pattern, it labels it as consumer type B. So it's make out of it machine basically which which label generate curry here. Label A, label B, label C in which we will put common pattern observe curry here. So there are hidden patterns, the machine is identifying those hidden patterns.

[00:01:49]One pattern for identify curry or which is the basis of which we will create different labels curry here. And this is basically your unsupervised machine learning.

[00:01:58]The third one that we have that is basically called as reinforcement learning in which we reinforce.

[00:02:03]Reinforce money we retrain, keep retraining. So it's like playing a game.

[00:02:07]We ask the machine to suppose I go ahead and I it's like building a chatbot also.

[00:02:12]So that is what machine company train karna shuru kiya. Then we reward the machine for every right answer and we penalize the machine for every wrong answer. And whenever we penalize it, whenever we give some penalty to the machine to which is the machine learns key you got it and the machine improves the next time. So it understands that something wrong was there and next time it will go ahead and answer the correct thing. So it's like in a play mode.

[00:02:36]Every time it does correct, we already give a reward. It does wrong, we penalize it and when it does wrong, then we because we have penalized it. You can use common punishment here. It will correct it in the second time. So this is called as reinforcement learning. I hope this is making sense to you and it is clear what is the difference between supervised, unsupervised and reinforcement learning. Okay.