Artificial Intelligence vs Machine Learning vs Deep Learning

Best Courses for Analytics:
---------------------------------------------------------------------------------------------------------

Best Courses for Programming:
---------------------------------------------------------------------------------------------------------

Best Courses for Machine Learning:
---------------------------------------------------------------------------------------------------------

Best Courses for Statistics:
---------------------------------------------------------------------------------------------------------

Best Courses for Big Data:
---------------------------------------------------------------------------------------------------------

More Courses:
---------------------------------------------------------------------------------------------------------

Full Disclosure:
Please note that I may earn a commission for purchases made at the above sites! I strongly believe in the material provided; I only recommend what I truly think is great. If you do choose to make purchases through these links; thank you for supporting the channel, it helps me make more free content like this!

[Eyedwiz]

Slight correction, deep learning is using neural networks which have multiple layers - hence the 'Deep' in Deep Learning. Otherwise it still falls into macbine learning as perceptrons

[AryanPatel-wbtp]

The key to deep learning is the activation function between layers you can have a neural network with no activation function that could be used similar to regression

[brandonmossop]

For ML you only described supervised learning, you left out unsupervised learning and RL

[pneumaofficial]

Anyone learning AI shouldnt be relying on shorts. Its an advanced subject that makes up more than just 3 categories and there are thousands of ways of going about each one

[perlkoblin]

In practice, it is just fancy terms to rebrand statistics

[hypermeero]

deep learning is using computational graphs to do machine learning, AKA neural network

so i guess the correct way to define deep learning is a computational graph.

[ShortFilmVD]

There's a difference between the media definition of AI and the academic definition, both of which have changed over the years. It's not too dissimilar to trying to define technology. For example, would you label a pocket calculator AI?

[johndododoe]

I as a human learned about AI when neural networks was an exotic and weak technique . Focus was on heuristic solution search methods like A* and rule based expert systems that had a true "why" command .

[dainionwest]

Another way of looking at ML vs. Deep Learning is in ML you manually engineer your features.

Using different levels of powers, combining features, splitting features that might have confounding variables etc. All of this you have to 1. Understand which variables to engineer and 2. Perform manually.

Deep learning does all this feature engineering for you, hence why a lot of applications even including image recognition, require a max of 3 - 5 layers of perceptrons before performance starts to level out (and why it's particularly useful for natural language processing)

[Rade]

Here you look most human. If you remove your specs you'll become rizz god.

[mk-ckor]

Can anyone give an example of something which is AI but not ML?

[musmustafa]

This is not exactly correct not all neural networks are deep networks. Deep learning has the characteristics of complex architectures with many layers

[CC-.]

AI includes all the different methods

ML includes methods of learning

Deep learning includes methods of simply learn very non linear data structures using nn

[sorvex]

Funny that statisticians try to claim all of machine learning by calling it statistical learning.

A lot of the progress is not even made by statisticians. It is like mathematicians saying that statistics is just Math.