A friendly introduction to Convolutional Neural Networks and Image Recognition

Announcement: New Book by Luis Serrano! Grokking Machine Learning. bit.ly/grokkingML
40% discount code: serranoyt
A friendly explanation of how computer recognize images, based on Convolutional Neural Networks.
All the math required is knowing how to add and subtract 1's. (Bonus if you know calculus, but not needed.)
For a brush up on Neural Networks, check out this video: • A friendly introductio...
For a code implementation, check out this repo:
github.com/luisguiserrano/man...
0:00 Introduction
0:22 Simple World
1:05 Keyboard
1:33 Image recognition software
4:39 Image Recognition Classifier
6:12 Artificial Intelligence
8:47 Gradient Descent
10:26 Slightly More Complex World
11:47 Previous Knowledge
24:27 Convolutional Neural Network
28:27 Advanced World

[@ravgeetdhillon2990]

Is this video eligible for an OSCAR? Outstanding.

[@deepakrkiran8471]

Best explanation of CNN ever found, Can you please create for LSTM and Autoencoder. A big thank you!

[@rekod_chanel3877]

totaly agree

[@SerranoAcademy]

Just uploaded one for RNNs, check it out! kzbin.info/www/bejne/i3_QooefhNOFnMk

[@NattapongPUN]

Agree

[@AbdulKalam-mf7wg]

Awesome sir Please discuss the parameters of each layer

[@basiratan702]

Great explanation with very helpful illustration... tqvm

[@PrevailingTechnology]

"Everything should be made as simple as possible, but not simpler." ~ Albert Einstein Bravo, sir, bravo! You indeed have a rare talent. It's difficult to distill down a complex topic into its barest essentials. More than a few highly-educated friends and coworkers look in utter bewilderment when they first encounter the topic. Too many "introductory" videos start by deriving the convolution function or with overly complex examples. Yours is a refreshingly "simple" and practical approach. Thank you and I hope to continue to share your talent.

[@itsnotallrainbowsandunicor1505]

It's depressing when professors fail to teach in a way that people can understand the basics with such clarity that students are forced to memorize a concept they just don't understand. The students continue never fully understanding concept, until wither it's revealed it through a co-worker, by reading/watching, or simply giving up and finding another venue.

[@naf7540]

Luis, I discovered your videos recently and I would like to thank you for the clarity and originality of your step by step explanations. Simple and close to reality at the same time, really great.

[@pushkarparanjpe]

Amazing level of clarity, use of diagrams and just enough redundancy to make sure the listener gets it! Great work!

[@mubashwir]

one of the best tutorials on CNN! Helped a lot. Keep it up.

[@shalinigovil7889]

What a brilliant teacher you are! I've been searching and searching for months to find a simple explanation of CNN that gives me the "glossary of terms" and I found you! Thank you!

[@winifredliang5683]

struggled with grasping the concept for so long, read so many articles and watched so many videos, i FINALLY GET IT NOW

[@supersql8406]

Thank you again, for your wonderful and superb teaching quality! Please do not stop making videos we need more quality instructors like you!

[@AshishSahu25]

You're a genius, whenever I tried to understand these, all I had was bunch of jargon, pretty hard to interpret. Please do it more, world needs you. You are the guy ai would like to hang out with. :)

[@salsabeelsh3878]

I've spent the past full week trying to figure CNN out from a gazillion tutorials and videos when all i ever needed was this one video! thank you!

[@SudipBishwakarma]

This is the best explanation I've found so far on the internet. The level of simplicity in your teaching is just exceptional. Greetings from 🇳🇵

[@MultiSciGeek]

This is so simple but so complicated at the same time. It's amazing what creative ways we have to figure all of this out! Really great presentation. Thank you very much :)

[@stevensong6909]

That was an amazing explanation. I feel like talking to someone about this. Its so logical and makes so much sense in why it works.

[@cardosct24]

Hands down one of the best explanations I have found on CNN. You have a gift for teaching. Some people know the material but can't transmit to others. You , on the other hand, know exactly how to break down the material!

[@ranga1179]

The intuition for how this is done was built up step by step and explained well in this video. Good to see the simplicity shine through.

[@davidmur7144]

Vow, what a wonderful explanation in a just half an hour. Really appreciate your very simplified guidance through such a complex topic!

[@RamkumarMambakkam]

You have the ability to explain in very simple terms. I enjoy seeing the video as i understand the basics easily. You should also do videos on statistics, data transformations

[@sanjayamv]

This single video is worth all the years of time that I have spent on KZbin .... Please upload more of such content...

[@naveenkumarp7896]

Your analogy by making our world and computer smaller was very impressive. You have the real ability to teach & transform. I would like to quote this here "A good teacher teach and Great teacher transform". You must do lot of videos to help students understand complex assumed technologies easier. You have that ability naturally. Thanks for the upload.

[@ajitpalsingh8200]

You are God's gift to Machine Learning Enthusiasts....Superb work....keep it coming....thanks

[@akhilchilumuru]

1. Trying to learn CNN 2. First video I looked at is this 3. Fully understood 4. Excellent explanation, Thank you Luis

[@rajasg2809]

Best tutorial so far... Subscribed within 4 minutes of playing the video.. Explained so gracefully and gradually that I didn't realize how 32 minutes of the video passed...

[@electrotsmishar]

Sir, you're such a great tutor. I've read your book "Grokking Deep Learning" and now I'm watching this video. Just can't express how much I liked your way of teaching....

[@lengooi6125]

Good stuff. Clarity of the explanation is superb. Thank you for the wonderful video

[@ShubhamSharma-vp6qn]

detailed explanation with simplicity, Thanks Luis Serrano 👍🏻

[@alessandrocagnola7605]

Thank you so much, this is the best intro I have ever listened to CNN. Simple but not simplistic, clear. Three minor suggestions I can give you for a possible 2.0 version are: - to expand a little the gradient descent. You calculated with patience the result of all filters but the gradient descent, in turn, is kind of evasive; - to complete the mapping to a neural network with weights and biases or at least give the idea how to; - explain the determination of the threshold (in the example you correctly put it to 3 but the determination is not evident). Conclusion: one of the best tutorial I randomly stumbled into. My sincere compliments.

[@akshatshrivastava4524]

Dude your explanation shattered my brain cells. Too good. Keep it up. Looking for more stuff from you.

[@MyselfBe19770322]

Thank yo so much. this is the best and most "friendly" one that I have seen so far.

[@nitinmms]

Wow!! Amazing. Have not seen a better ex-plainer for CNNs. Thank You.

[@erhanakan3298]

That was the most friendly and simplest tutorial I've ever watched. Thank you.

[@naidu19901]

Such an intuitive way to explain complex CNN with simple / and \ . I'm glad and fortunate to found this tutorial while searching for other videos. Thanks much for great effort and sharing publicly with us. Hope to see similar kind of great explanations for other Complex problems of Data Science from you. !!!!!! Cheers !

[@shyambutani8618]

It's the best introductory explanation of cnn. thank you very much.

[@Icglez1921]

Truly a video for beginners. Nice and Simple. Thank you :)

[@kamilawhocares8362]

I've seen some explanations of CNN but this is without a doubt the best one I've come across. Thank you very much for this video. So well done, so well explained.

[@bienne777]

His voice was clear and presentation was clear to me to learn about the basics of the topic, making things simple and easy to understand the concepts.

[@phaniauce]

Thanks for this detailed explanation...You explained very well.

[@shivipatel9634]

One of the best explanation of cnn..... Hats off to your way of explaining the concepts..

[@vatssv3006]

Going to save it to favorites because I know whenever I will have doubts about the clarity of my concepts I will revisit this and share it with others. That's how good this is.

[@prasannadevulapalli9466]

Simple & clearly explained! I felt it is a very complex concept when I read the papers on this, but I feel its very simple after watching this video. Kudos to you sir !!

[@banyantree]

EXCELLENT INSTRUCTOR!! Well organized progressive explanations. Well presented. Good speaking pace too. Would love to see more instructional videos from Luis. THANKS!

[@praveenraghuvanshi6489]

Great Explanation... you just un-convolved the hidden layers of a complex topic that was revolving in my mind :-) Thank you

[@shoogarh6797]

You are the best instructor in the world. Thank you so much for you amazing work. You give life.

[@shashankgarg3258]

Hello Luis, Your courses on CNNs and the related teaching methodology are truly fascinating and brilliant! Thank you!

[@ThorNjord]

Another awesome video! You are the only I found who really explains the intuition behind the math really well! What do you use to build your slides?

[@CodexAudentia]

+1: I want to know that too.

[@SerranoAcademy]

Thank you, glad you liked it!! I use Keynote for the slides.

[@ThorNjord]

You have so many cool animations, I thought you had found something better than Keynote. :) Thanks! The way you do them really helps to understand so keep up the good work.

[@SerranoAcademy]

Thanks! yeah I just spend a lot of time on them. :) I need to do some 3D ones that are harder on keynote, so if you know of any good 3D animation software let me know.

[@shivasthilak]

blender is free and offers a lot off the shelf. Also,a question for you. If I want to build a prediction model from multiple data sets(at least 4), what is the best way to choose the right classifier?

[@shafiudinsyedmohammad2919]

Hello Luis, This is Shafi, I watched your videos on Machine Learning and Deep Learning, I am very glad to say that It is a great work to explain ML and DL Algorithms in a simple manner so that even a lay man can easily understand what it is, it helps a log to dig into ML and DL. I would like to suggest you to please provide one lecture for Recurrent Neural networks. Thanks a lot for your work . Many Thanks Shafi

[@SerranoAcademy]

took your advice! kzbin.info/www/bejne/i3_QooefhNOFnMk

[@superasn]

Very good explanation. You never once assumed that anything was abvious to the viewer. Great channel and video, deserves a lot more views!

[@anummalik7733]

I dont know how u explain this all so easily.. appreciated. Your concepts are very clear.. keep it up

[@thejishubasak]

27:38, that moment when you will get a mini heart attack :O

[@vatssv3006]

I literally thought my screen is gone. To top this, I dropped my laptop a couple months ago and it was the screen that was damaged then too!

[@hackercop]

I thought my GPU was having a heart attack lol

[@afsangujarati9427]

23:01 That sounded like a Plus One song! :D Amazin explanation, watching it for the third time.

[@felicia6242]

I love how simple you made to understand basic concepts! Thanks for sharing!

[@cspineda]

I am currently working on a presentation of CNN and was confused about the actual way the convolutional and pooling layers worked before watching this video. This video really helped me understand the basics of the ConvNet layers. Highly recommend!

[@lotfichikh9417]

Very pedagogic explanation of CNN ... congratulations for this work !

[@colleenkuang7778]

Awesome video!!! It gives the insight of CNN. After watching it, I completely understand how CNN works.

[@myweg7806]

I loved it. You did an amazing job and saved hours for many people who would be going through many sources before understanding the concept clearly, spending a lot of time. Thanks a lot.

[@onebeartoe]

Oh my goodness, this video on convolutional neural networks help me understand them so much! Thank you, Luis!

[@qwmp]

Great! This is one of the best explanation of convolutional nn!

[@SerranoAcademy]

Thanks Sungjin!

[@sallerc]

Another great one, thanks Luis!

[@SerranoAcademy]

Thanks! :)

[@javiercarrillomartinez1289]

Wow, I really love the simplicity and clearness of your explanation.

[@sourbhopal]

I was looking for this kind of video for that last one month ...you made my day..simply awesome

[@Mobin92]

Aren't convolution and pooling actually very similar? How is pooling different from a convolution with just one filter?

[@victorarias9688]

owwww shit, this is the best explication in the world!!!

[@paedrufernando2351]

Ultimate intuitive series. Thanks for the Knowledge sharing..I think I am able to understand it now..Also parallelly I am learning the math of it through the other courses..so able to connect the dots

[@user-bp5yo6jv6e]

Great video, It's amazing how you managed to take on a mathematically complex subject, And by simple examples simplified the complex theory Thank you!

[@mayankverma4271]

I think the explanation is good , but I feel one problem in this video about Convolution and Pooling explanation . How to choose filters and how much with no clear explanation about pooling . Rest is good and again thanks to sharing .

[@itsmemouha2811]

This is one of best introductions I found on convolutional neural networks! Thank you so much!

[@hari20001]

Hi Luis, I agree. Fantastic explanation. Thank you for making this and your time. You are a great teacher!

[@FlavioChianese]

The BEST and EASIER explanation about "CONVOLUTION NEURAL LAYERS", my compliments! Greetings from Italy.

[@freedomtweezle8697]

Luis, awesome, awesome job. Builds from the base case, summarizes, and shows concrete examples. Bravo. More, please.

[@andrewchen861]

Awesome video! I have read a few articles about convolutional neural network, but this video really helps me understand! Thank you very much!

[@dompower500]

There are many videos are made by different experts by their hard work on this subject. However,Your teaching and examples are the best so for I have seen. Thank you sir

[@blesucation4417]

Just want to leave a comment so that more people could learn from your amazing videos! Many thanks for the wonderful and fun creation!!!

[@ivogeorg]

It is interesting how long it took between the original Yann LeCun paper (1999), the foundational scientific publication on convolutional networks, and this marvelously distilled instructional video (2017). This pattern can be seen throughout science and technology: some domain expert formulates the theory in its full glory and complexity, and then someone else takes it apart and reduces it to the most basic and intuitive patterns for the instruction of novices. These two works are usually produced through very different mental labor and are rarely the work of the same person. One notable exception is Richard Feynman, whose fundamental breakthroughs in quantum physics and lectures on physics are equally lofty pinnacles of science and education, respectively. Kudos and thanks, Luis!

[@anupgogoi1402]

Hi Luis, your friendly videos not only make understanding and career but also it makes life. Thank you so much for this amazing video. I appreciate for your effort. God bless you

[@jaredfiacco1107]

Incredible video, you can tell how much time went into this. It really made a big difference on my end! Thank you!

[@visualsbysri]

Very crystal clear explanation, helped me a lot to remove any confusion while doing masters!! Much thanks!

[@sanjaykrish8719]

Awesome. This is the first time I am enlightened by seeing a video in neural network. Thanks a lot.

[@finkga]

Luis, This is awesome. I understood how to use CNNs before, but now I _really_ feel that I understand the reasoning behind them. Thank you so much. You definitely have a gift for teaching!

[@peymanmortazavi5054]

This is by far the best tutorial I've seen, kudos to you Luis!

[@tyfoodsforthought]

For sure the best CNN diagram and explanation I've seen thus far. Thank you!

[@bordgame]

really really amazing I couldn't imagine someone can explain a complex concept so simply and also completely

[@madhubagroy]

Luis, thank you so much. You are simply THE BEST!! I have benefitted immensely from your intuitive explanations (including the Udacity ND ones)- it's a joy!!

[@daniel5992]

Wow, such an amazing video! And finally one, with some understandable examples from beginning till the end! A BIG high on you!

[@TheAkshatJ]

One of the most apt videos on the topic of neural networks, on youtube. Very lucid way explaining the concept. I almost started laughing at the beginning when the world with two characters and two by two pixel PC was introduced. Hats off. Looking forward to next set of videos...

[@PeterPan-hs5tu]

this is by far the best Barebone illustration that I’ve seen and easy to understand the concept of CNN, bravo!

[@sarathsg1964]

That was some truly brilliant explanation. Simple and precise. Amazing talent to explain a concept and make it look so simple. Simply awesome !!

[@chandnishah1780]

Very well explained! Thanks for taking time to narrate this so well with all the examples!

[@sayajujur2565]

Salute to Luis, one can give simple explanation of complex problems only after thoroughly understands it. Udacity should be glad to has Luis.

[@papangbuamanee9571]

I've been studying for weeks and this is the best one, easy to understand!! Thank you very much!!

[@arminkashani5695]

Simple as it is, this is truly a masterpiece. You have made it so straightforward and intuitive. Thank you.

[@lazaradamovic2813]

One of the best if not the best video explaining CNN's online! Bravo!

[@lohithArcot]

Wow. I never came across a trainer so articulate. I loved when he said how do we recognise these images using maths

[@dr.rudranarrayanbaral9595]

Never ever taught like this. Great teacher with the best knowledge.

[@akanksharathore3946]

Thanks to this tutorial I finally understood the very basic concepts behind CNNS. Thanks a lot Luis Serrano!

[@adrianorcampos]

Incredibly awesome video! I agree with Deepak when he said that this is the best explanation EVER!!!

[@adamchace]

Excellent explanation, clarified a lot of concepts that were hazy from other sources.

[@shailatabassum5360]

I couldn't realize properly after watching many videos on CNN but really i just enjoyed to watch it.. You are a very good teacher ! Thank you! 😊😊

[@addcolor8326]

Simply amazing intuitive explanation.Far better for beginners than Andrew,Karapathy et al.. Thanks for putting the lesson on KZbin.

[@lifehacks4everyone]

this was the simplest explanation that i have come across.. thanks a ton Luis!!

[@MrBlueyes61]

Excelente explicación, Luis. La más clara que haya visto en años. Gracias por compartirla.