Derivation of Backpropagation in Neural Network

Understand the Maths behind Backpropagation in Neural Networks. In this video, we will derive the equations for the Back Propagation in Neural Networks.

In this video, we are using using binary classification. In the next video, we will see the Backpropagation in Neural Networks with for mulit-class classification

This video, will be superhelpful for implementing Neural Networks in python.

Timestamps:
0:00 - Video Agenda
1:30 - Important Shapes and Equations
4:25 - Derivative of Loss function w.r.t W3
12:00 - Derivative of Loss function w.r.t W2
16:08 - Derivative of Loss function w.r.t W1
18:33 - Summary

➖➖➖➖➖➖➖➖➖➖➖➖➖➖➖

This is Your Lane to Machine Learning ⭐

➖➖➖➖➖➖➖➖➖➖➖➖➖➖➖

➖➖➖➖➖➖➖➖➖➖➖➖➖➖➖

[MachineLearningWithJay]

If you found this video helpful, then hit the *_like_* button👍, and don't forget to *_subscribe_* ▶ to my channel as I upload a new Machine Learning Tutorial every week.

[mdtufajjalhossain]

I am mesmerized by the way you teach. I love this. I find learning NN difficult but after watching your videos and explanation, it seems easy to me. Thanks a lot. Please keep uploading your good work, no matter how many views your videos get. I believe, your videos would certainly be popular once people notice them.

[ranjeetkumbhar]

Seach a beautiful derivation with simple and minimal use of notations!!!

[PrithaMajumder]

Thanks a lot for This Amazing Introductory Lecture 😊
Lecture - 5 Completed from This Neural Network Playlist

[kingcrimson]

excellent work here! I was having trouble coming from single neuron in each layer where the matrix multiplication was not matching. Thanks for being so clear on how to manage the transpose on the weight function! that was exactly the missing info I needed!

[vicdata]

The best worked out example I've found. Good stuff.

[abhishekchaudhary]

Ur video are so good I revisit them every month.

[vont]

Thank you!! These are super informative.

[kevinhoward]

Loving this series. I could be mistaken, but I believe in the summary slide, the equations for operations 'dW2' and 'dW1' are incorrect, as the 'DZ2' operator should be multiplied with 'A1T' and the 'DZ1' operation should be multiplied with 'XT'. Didn't hinder at all, just wanted to point it out. Thanks so much again. 👍🏻👍🏻

[simpleman]

I love this video. Thank you so much for explaining such a complex concept so simply

[abhishekfnu]

Thank you so much for the explanation bro. This video means a lot to me. 🙏🙏

[garimakaushik]

great help. Thanks for all the videos. too good

[aniketjalsakare]

Very helpful, solved my little doubt
Keep it up bro

[GK-jwbn]

thank you so much! very helpful! backpropagation was a hard nut to crack. NOT ANYMORE😛

[aloktrivedi]

Great Lecture. Clear and concise . Keep up good work

[vaibhavoutat]

At 14:38 you say we can take the element wise multiplication, I see the dimensions are same but how did you know we have to take element wise multiplication or a matrix multiplication like we did previously for getting delw3. I am looking for an answer that does not just say that we need to match the dimensions that's why

[christopherwashington]

Thanks for the tutorial! Is there a theoretical back up for the part that you make up the correct dimension?

[himtyagi]

Excellent clips..thanks.. I think there is error in last slide..dW2 and dW1 will depend on A1 and X respectively..

[siddharthc]

Could you explain why we take element wise multiplication at 14:46 and not regular matrix multiplication?

[sivaprasad-clxp]

Can you please explain or share the documentation by taking cost function as mean squared error and perform for regression?