Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM)

[@brettensor]

I really appreciate how you wait to the very end to discuss the math of RNNs. Your approach of leading with the intuition of RNNs rather than the math was super helpful for me. Thank you for your work!

[@BrandonRohrer]

Thank you :)

[@aryanr7490]

7 years ago video, 9 months ago comment. It's so cute gesture that you still go back to your old videos @BrandonRohrer .

[@ssundaraju]

Brandon has applied extraordinary skills in communicating the difficult (convoluted?) topic and concepts on LSTM, in simple and comprehensible language. The example on writing the "Children Text Book" brings out all of the major processes into sharp focus. I have learnt immensely. I will now be studying other videos by Brandon. Thanks a lot.

[@tomforstmann9977]

Kind of meta-ironic watching this with automated subtitles enabled. Really good video on the subject, much appreciated!

[@BrandonRohrer]

:) Love it

[@kerbyboy43]

im a student major in data science from taiwan and i wanna say thank you, i got more to know about LSTM after watching ur video. Much appreciate sir!

[@BrandonRohrer]

You are very welcome. Best of luck in your studies!

[@gergerger53]

It makes me want to make four more KZbin accounts just to give the video more likes. I've been staring at the network diagrams on blog posts for a few weeks now with no idea how to really read it. That simple walkthrough on the small example was just perfect.

[@Spudst3r]

Your videos are absolutely amazing. You are the first educator to make these complex topics truly accessible to the general coder without the privilege of getting to attend Stanford. Thank you so much. Everyone I've shared this to has just eaten it up as well.

[@Neonb88]

I actually thought there was a fly on my screen. Kudos!

[@StarAI24]

@@Neonb88 didnt see the fly until you mentioned it lol

[@georgiostsoukas7270]

Just brilliant. Not only learned about LSTMs but also how to do a proper presentation.

[@BrandonRohrer]

Thanks Georgios!

[@jordanlee1439]

The pace and clarity of the lecture is really refreshing. Methodically stated complete sentences are always welcomed in my book.

[@GEB-Loop]

The best explanation of RNN and LSTM that I have ever come across. Thanks for sharing!

[@BrandonRohrer]

I'm happy you liked it!

[@Skydmig]

Everyone give this man an upvote. This is great material.

[@asteroidX]

An upvote doesn’t justify. Plus hit the subscribe button.

[@carlseconds7557]

I don't usually place any comment like this, but this is extraordinary :) So easy to understand :) Thank you :)

[@BrandonRohrer]

Thank you very much!

[@rainbain5474]

I have been getting into neural networks for a while now, and this is the best material I have seen.

[@aaryannakhat1842]

This is probably the best video on RNN and LSTM that I have come across!

[@BrandonRohrer]

Thanks Aaryan :)

[@hessamjamalkhah9781]

Dude you nailed it, this simplicity on explaining such a complex concept shows how great you are and the depth of your knowledge on the subject. Thank you so much.

[@BrandonRohrer]

Thanks Hessam, that is a giant compliment. I'm glad it helped.

[@Tfrexbex]

Hollllyyyyy cow. Thank you. I have only just started this vid but as a TBI survivor who taught myself how to converse due to recall and memory issues along with researching the SCIENCE…. Can’t tell you how invaluable this validation is.

[@AurL_69]

This channel is a goldmine ty

[@sqripter256]

Easily the most intuitive intro tutorial on RNN/LSTMs. Great work and thanks

[@swethanaik7121]

A very clear and concise concept of LSTM. Had a hard time trying to understand LSTM until I fell upon this video. Thanks Brandon for this

[@BrandonRohrer]

Thanks Swetha! I'm happy it helped :)

[@francescocalifano4856]

I'm preparing my university exam about "Data Science and machine learning", these videos are pure gold for someone who is approaching this topics for the first time like me. Thank you so much, it was really worth the time spent to watch it.

[@mannversteckter5631]

The example with Writing story is very clear. Thank u!!

[@fgfanta]

The most approachable introduction to LSTM I have found so far.

[@DreamWorker-jm5xn]

Yes, this is the best LSTM video on KZbin. Today is 27th July 2019.

[@dattatreyapujar4068]

What is the day today?

[@bode_fuceta]

27th July 2020 *(1 year later exactly)* ...Best video yet

[@tasteyfoood]

As someone who learns best through story and images, this has been incredibly helpful. Thank you!

[@TheodorosKatsikis]

This has to be the best LSTM explanation I have seen! Awesome work! THANK YOU for making this

[@jaydayyyy611]

I have never seen this level of simple explanation of LSTM ever. Awesome..

[@kartikpodugu]

Saw many videos on explaining CNN in a simple way. This video explains RNNs and LSTMs in a simple way. Very intuitive

[@Lines-In-The-Sand1]

This is by far the best channel out there on machine learning. More please!

[@Vladeeer]

agreed

[@ajitkumar15]

This is really amazing video on LSTM. Pace and clarity of explanation is awsome !!! thank you Brandon

[@BrandonRohrer]

Many thanks Ajit.

[@011azr]

You explain everything in a crisply crystal clear way. What a great teacher! :)

[@garrettosborne4364]

Great job Brandon, simple and straightforward explanation easy to grasp concepts.

[@BrandonRohrer]

Thanks Garrett!

[@fuuman5]

Insane explanation. Just invested 26 minutes of my time perfectly.

[@eashansoni3761]

Damn, the people who come up with these neural network architectures are brilliant.

[@balmukund6181]

Thanks so much Brandon!! I wish to learn how features are extracted from network packets. Thanks for the videos. Urge all to give you an upvote.😀

[@BrandonRohrer]

Thanks Balmukund! I'll keep network packets in mind as an application.

[@balmukund6181]

Thanks a lot for your reply Brandon! I am trying to use CNN algorithm for network security application for my project in my masters. Could I connect to you Sir via mail?

[@vyaslkv]

I don't usually comment, but Mr Brandon the way you explained things really made to do this. Really superb sir _/\_

[@goodmusic284]

The best Intro To RNNS and LSTMs I have seen!

[@Jabrils]

Hey thanks for this video brandon! I learned a lot of great new LSTM knowledge. I am coming over from Luis by the way 👍

[@NickKartha]

Which Luis, Jabs?

[@roniquinonez9715]

Luis Serrano. LUis Serrano's Friendly Intro to RNNs: kzbin.info/www/bejne/i3_QooefhNOFnMk

[@BrandonRohrer]

Overdue on my part, but I'm so happy you found it! Thanks for the shoutout. Keep up the amazing content. You are doing great things.

[@Leonardo-fm7fj]

@Karsyn Denver Why on earth would you like to hack your gfs instagram? What are you, a thief or something? At which country do you live in? I am sure the local police authorities would like to have a word with you or two.

[@NositisNation]

Just excellent. It is art to find the simple story w/n complexity. Well done sir, well done.

[@PradyumnSharma]

Brilliant! Such lively, easy-to-understand explanation. Wish there were more teachers like you.

[@sanjibphukan3570]

Beautifully well illustrated in simple, step by step manner for such complex concepts

[@DanTaninecz]

Astonishingly great video. As an instructor, you provide a lot to emulate here.

[@japsowinkaur525]

The best video on LSTM and RNN I've seen. Thank you so much!

[@stevecoxiscool]

Fantastic explanation. Taking the time to explain the diagram symbols really helps (me) (sigmoid/tanh/addition/multi). I also liked how the video explains that with RNN/LSTMs, you are going to need several NNs in this solution and not just one NN with lots of hidden layers (another words what DEEP really means). The other hard part for me anyways has always been data conversion, knowing that you have to obviously convert a "string name" to a numerical representation (TF/IDF) for example before you can even push the input data into a NN architecture. Thanks for your effort !!!

[@swathykrishna9618]

Thanks. Well explained. Finally a nice explanation for LSTM s.

[@BrandonRohrer]

Thank you Swarthy :)

[@nguyenchau3800]

Wow, I've been looking for this kind of explanation about LSTM in details for weeks. Thank you so much!

[@sebastianappelt4652]

The best explanations for CNNs and RNNs! Thank you a lot!

[@philippemuller3079]

This is simply amazing and so clear.

[@vijayakumar4804]

Hey Thanks Bradon, Complexity put forth in Layman's language, just loved it!!!

[@huyhoang752]

The best introduction to RNN and LSTM on KZbin! Thank you, Brandon!

[@khallild2288]

I watched it 3 times and I finally get it, it's really a fantastic system. Thanks for the really good explaination

[@jaimegonzalezsuarez1566]

THE BEST VIDEO EXPLAINING NEURAL NETWORKS BY THE FAR you a re and amazing teacher, everything is clear you have a new fan good work

[@hitashukanjani4430]

Best possible explanation of LSTM

[@normansabin8322]

I love the birds chirping in the background, super calming :)

[@hmoazed]

Thanks for simplifying something complex so clearly that even I understood it :)

[@pallavijog912]

This is an unbelievably amazing video that I have ever seen for RNN/LSTM.. I was looking out for exact systematic approach to explain it which is not done by anyone else.. Subscribing to your channel.. Kudos!

[@robindong3802]

this is first LSTM video, you gave me a confidence to go on. Thanks,

[@rembautimes8808]

Thanks for the video. I started on ML 2 months ago (Dec 18) and enrolled in Coursera's Machine Learning course. I liked the way how you presented the material. Even though the material is for language processing, I can see how this could apply to financial markets. Example if certain technical indicators certain to occur in sequences. The forgetting function is hopefully useful to sell at the top and buy at the bottom. Btw, my research interest is to develop a "language" for technical indicators, not the indicators perse how they occur in sequences. E.g. TTM Squeeze, followed by Doji, etc.

[@Damodharanjay]

Hands Down! The best LSTM video I have ever seen! Thank you!

[@DagarCoH]

I have seen a couple of videos on the topic, and yours is easily the best. I would have had no chance to understand the other ones, had I not watches this here first. Thanks a lot!

[@ibrahimelkilani8508]

Really. Other videos do explain in hard scientific way unlike this one

[@mihkellukats4773]

Best graphical explanation I've seen so far. Thanks!

[@easylearn9350]

I am beginner in DNN, I must say, I really enjoyed learning LSTM, Thanks a lot!

[@BoruffAdventures]

This made LSTM clear and understandable. Also your How Deep Neural Networks Video made the abstract way a computer hones its network down a simple and clear algebra equation.

[@jackshaak]

Great introduction to LSTM. Thank you sir.

[@AnuragSingh-xi8kw]

Great tutorial being a beginner i learned almost everything..

[@BrandonRohrer]

Glad to hear it!

[@daljeetsodhi670]

Hi Brandon, you made my day today. Very nicely and simply explained

[@ashlife17]

Absolutely wonderful and simplified video explaining LSTM. Thanks a ton Brandon!

[@piers9186]

Clear, unpretentious, non-shouty video. Thank you.

[@yysatoshi]

Is it only me who has no idea of what this genius is talking about

[@saint79209]

😂😂😂 good one

[@alignedbyprinciple]

You are such a great human being dude. You make this complex concept very easy to grasp even by those who think Math is very scary subject. I have no clue how easier can those paid companies can explain this. Hats off for you sir. Education SHOULD be free, it is human right. I will do everything in my power to share my little knowledge to the world. Let's make the world better by destroying greedy money hungry folks out of education. I would appreciate it if you can explain the concept of state in this context. What is it exactly representing?

[@thewho2801]

Thanks a lot for this straight to the point explanation. Really learnt a lot.

[@RrTalks4320]

the explanaation given for LSTM is simply superb. You have told, 3 mistakes in RNN. 1) Daug saw Daug. 2) Jane saw spot saw Daug saw jane. 3) Spot.Jane.Drug.. Problem 1 is solved using, forget gate to reject the same ip cannot be a o/p as drug. Where other 2 problems in your lecture.

[@helkin96]

Same idea. The forget gate will have "not saw" saved after Spot and "not ." saved after Jane. Therefore, those options will be cancelled out, just like Doug was in his example :)

[@Vikram-wx4hg]

Thanks for taking the pains to build those examples. You are explaining this wonderfully!

[@georgplaz]

There are so few good videos on this topic and you nailed it. thanks!

[@Apurvsankhyadhar]

Absolutely amazing video, especially the toy example at the end. Was struggling with Andrew's video, but combining the intuition from this video with the diagram and formulae he told, damn am I happy

[@eunhyoukshin7777]

I'm an undergraduate student and your videos are so helpful and encouraging! Thank you for the amazing video lectures :)

[@stayinthepursuit8427]

Did you graduate

[@majeedhussain3276]

This is the best explanation on Internet

[@slimyelow]

This is an extremely helpful and productive explanation plus creative demonstration of the RNN. Thanks so much !

[@ashita1130]

Watching in 2x speed, even then it makes perfect sense! 🙏

[@theGreaterAwareness]

It's interesting. I once got high and came up with the exact same idea for AI but I had different terms for the components. Like, 'forgetting' I called compression. Vectors I would call snapshots. This was like 4 years before I even knew what a neural network was. So it's like relearning what I already know but learning the accepted terminology for it. I also did the same thing when I taught myself music. I had my own terminology for chords and scales and then I made friends with a musician and he taught me all the terminology for everything I already knew. So it's strange, it's like having someone tell you about your favorite hobby in an alien language and making it sound more difficult than it is - well, your video is great, probably the most easy and natural sounding video I have seen so far but some of my fellow Aspies....it's like they go to university to study how to make things sound as awkward as possible and too complex. Example: "Now if you take your doppleganger tetrahedron of mu and multiply the dexarangatron by the valamucci of scrabaducci then you get the square root of bigawum which is so obvious that if you cannot keep up with my monotonous voice then you should watch my introductory video where I tell you the most obvious things that will not help you at all, and where you can buy my book: An introduction to how I make things sound as unappealing as possible". Thank you for your video, it's like riding a bike, looking at the sunset and learning about something I enjoy. Now I'll be able to keep up with other people whom are interested in the same thing because I'll know the terminology they are using.

[@ravisawhney3111]

Great intro to RNN and LSTM for a non-developer. Will follow you.

[@shridharyadav88]

That was an amazing explanation. Very clear with the help of images and symbols.

[@piotrekgor18]

after just 3 minutes i knew it is going to be a brilliant video. Thank you.

[@YH-fj9bs]

This video is the best one explaining the LSTM!!! Thanks you so much!!!

[@allisswellable]

Excellent explanation of LSTM in a very simplified and realistic way... Great video!

[@arslanullahpractcingphyics8969]

you really deserve big hand. such a fantastic explanation you made my day sir

[@ardeshir1912]

Great Video explaining why and how we had to modify the simple NN to achieve this. Thank you. Looking forward to more videos

[@realcolby]

Thank you so much I’ve been trying to understand LSTMs all week and this video finally did it for me

[@zvzenjoyer-ao]

Best tutorial on RNN and LSTM !!! Keep going !! I subscribed after 4 mins ! :D

[@ramalingeswararaobhavaraju5813]

Thank you Mr.Brandon Rohrer sir, good evening sir, thank you Mr.Brandon Rohrer sir for your good teaching on Recurrent Neural Networks and Long Short - Term Memory (LSTM).

[@Kevin-cy2dr]

Dude you dont have to call him sir. I bet he's just ok with brandon

[@daulatrachmanto2353]

Tbh, this is the best video explanation about LSTM i've ever seen Thanks :)

[@AbhishekGupta-zx9dn]

By far the best explanation on this topic.

[@frankli5075]

Best one to explain LSTM so clearly.

[@baskaisimkalmamisti]

I can't think any better explanation. Wonderful!

[@ocean694]

Exceptionally good, the best I've seen in this subject.

[@qianren7207]

It's excellent to show the Long Short-Term Memory by data flow rather than mathematical formula.

[@TayiAbhignyaBCE]

Thank you so much, Brandon!! It is a great video. I understood the concept really well.

[@BrandonRohrer]

That's awesome! I'm so glad.

[@smokedchickendrum]

Awesome video indeed. Very abstract concepts explained in very straightforward visual figures. Really appreciate it.

[@saurabhvishwakarma8761]

Extremely helpful video, your explanation is very clear and to the point

[@gelliravikumar018]

Apreciated Brandon, Nice presentation. Crisp and precise explanation for LSTM concept.

[@MauiMana]

Excellent walk through of fundamental LSTM concepts

[@goody4177]

Really nice illustrations and examples! love it :)

[@BrandonRohrer]

Thanks Goody!