วีดีโอ

8:11 Hypothesis Representation

25:20 Math behind large margin classification 45:04 Kernals I 1:00:49 Kernals II

0:15 Model Representation 11:15 Cost Function 16:25 Cost Function Intuition

very helpfull

Nice class

here after the 2024 nobel prize announcement in Physics

Nobel Prize

This is the greatest introduction of ML

Deep respect! Thank you for this beautiful lecture!

very much helpful ❤

I remember walking into the Spice Rack and seeing the (iirc) Perq machines running boltzmann machines. Or maybe they were Symbolics machines?

37:09 Theta_0=0 does not mean the decision boundary goes through the origin... Theta as drawn on the left has positive first entry...

props to you :v

I feel like Prof. Hinton is trying to avoid math to fit the understanding level of coursera users. However, instead of conveying the common sense and insights behind the functions, he simply rephrase the functions into everyday language. Since everyday language is extremely vague and unclear, he created difficulty for even math lovers to understand the ideas behind. I highly value the ideas Prof. Hinton tries to convey, but to really understand the ideasis hard. Thanks for sharing anyway.

Hi Blitz kim, may you please assist me in how i can add loads to an octave portal frame?

Thanks a lot

Thanks a lot

00:00 1A Why do we need machine learning? 13:15 1B What are neural networks? 21:45 1C Some simple models of neurons 30:09 1D A simple example of learning 35:47 1E Three types of learning

This channel is sooo educational. Thank you!!

I look up to you sir

wow!

Hi. Anyone who understands what we mean by a training case? I do understand what an input vector is, but not a training case. Help would be appreciated.

Any one who can explain how the network is learning the weights in the hand-written digits example? I don't get the increment and decrement weights idea.

Inspirational story about importance of Being a balanced controller th-cam.com/video/4BiKhka7APc/w-d-xo.html

This is the best video on hopfield and boltzmann machines on youtube.

I am surprised that I am first to like the video. This is so nicely explained.

you are awosome

What a brilliant explanation! Thanks a bunch.

Four effective ways to learn an RNN: 35:24

Very insightful

I just wish he would define his terminology a bit. E.g. “greedy”, etc. But otherwise, brilliant. 31:05 probably how the brain works.

Dense and informative lecture, but starts to make sense after repeated viewing.

Learning from the master. Unfortunately, he doesn’t explain at all about the “convolution” part of CNNs.

This guys knows what he's talking about.

What is mean by input vector? Why input vector is perpendicular to the hyperplane ? how to differentiate input vector and training case hyperplane?

This man is The Godfather of AI...really genius

30:10, 'the backward pass is completely linear' hahha oh man that was hilarious!

"This was really amazing in the year 16 BG" XD XD XD XD

The GOD of Neural Networks ( Geoffrey Hinton)

truly awesome! So many insights.

He is a brilliant explainer

Great!! Thank you so much

What about fitting the model (hyper-parameters) to the training set, rather than the validation or test set? In other words, taking your trained models, and see which one performs best on the training data. Is that an acceptable practice? Does using the validation set reduce over-fitting to the rest of the training data?

SVM math is a challenged me, but I am determined to master it.

So it all comes down to the most frequently occurring "n-words"

many thanks for uploading this

19:00 what the hell

Possibly one one of the more concise and complete explanations on how to visualize the process... Side note, it looks like `Raghavan` of `Medium.com` agrees because they totally took a screen-cap medium.com/datadriveninvestor/feature-scaling-why-what-where-how-683f61812f4c though without citing a source. ... though much like +Nurhan Serin (however, not for the same reasons) I too think a bit of time spent on how reversing things can be done would be super helpful for those just getting into the subject.

I want to ask how to denormalize what we found h function. Because I think it is a neccessity to transform normalized h and y values.

Does anybody have the link to where I can get my password?