AI - Artificial Intelligence:
Subsets:
1. Machine learning
2. Deep learning
3. Data science
4. Computer vision
5. Robotics, etc..,
To get a detailed understanding on this refer the below image:
Machine learning:
Machine learning is the process of feeding data to machine. The machine learns by using those data and thus becomes artificially intelligence. There are some types based on the way we train the model.
1. Supervised learning:
Machine learning that learns with labelled data.
Eg: Consider face unlock in your mobile phone. You should register your face by taking snaps by using that phone. Your phone stored your information by labelling those photos as the phone's owner. So that when it find the phone's owner takes the phone it recognises that it was the phones owner and it unlocks the phone. When other people take the phone and show their face it recognises that the face doesn't match the phone owner's face. So the screen doesn't unlock.
2. Unsupervised learning:
Machine learning that learns with a unlabelled data. The machine just learns its characteristics and tries to differentiate between those characteristics.
Eg: Consider cricket players for this example. Take example of Jasprit(bowler) and Kohli(batsman). Now think that the machine has been fed with data of their previous One day international match scores. The machine analysed the score and return bowler if the player got more wickets, if the scored runs was higher then the machine returns batsman/
3. Reinforcement learning (Semi supervised learning):
Machine learning that is achieved by continuous feedback and experience.
Eg: Lets consider a robots task is currently in a classroom and its task is to get out of the class. In order to completed its tasks the robot first stand up and tries going straight. But it faced black board and found that the exit is not in the straight, then the robot tried going left side. Again it faced some desks and found that there is no exit way on left. Finally the robot tried moving right and now it founds that the exit is on the right side. So successfully the robot got trained and it used its previous experience to get out of the class easily.
Bonus point:
You should also know this:
Symbolic learning: Symbolic learning is a traditional way also known as symbolic reasoning. It is also considered as an subset of Machine learning. It uses symbols to represent certain objects and concepts, and allows developers to define relationships between them explicitly. Symbolic learning lost its popularity due to growth of neural networks.
By this hope you got a basic idea about AI. Happy learning :)
Comments
Post a Comment