Variational Autoencoders | Generative AI Animated

Поделиться
HTML-код
  • Опубликовано: 19 дек 2024

Комментарии • 145

  • @essentiallifter
    @essentiallifter 3 месяца назад +50

    this account is seriously underrated, will definitely blow up soon

  • @Higgsinophysics
    @Higgsinophysics 3 месяца назад +20

    I can't believe how calmly and clear you explain difficult topics!

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад +13

      Well that's the magic of text to speech 😁

    • @viewer8221
      @viewer8221 2 месяца назад +3

      haha I knew it was AI voice

  • @AtiqurRahman-gj6mg
    @AtiqurRahman-gj6mg 3 месяца назад +13

    Finally, the concept of VAE is clear. Thanks a ton.

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад +1

      You're welcome, thanks for the comment !

  • @xichengwang1319
    @xichengwang1319 2 месяца назад +5

    I first saw this in Chinese media with Chinese subtitles, then just came back to subscribe to the original author. The most clear introduction ever seen with such nice proper animation. It will blow up for sure.

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад +1

      Well thank you ! Can you send me more info about that through my email ? ytdeepia@gmail.com

  • @anthonyortega358
    @anthonyortega358 3 месяца назад +7

    This channel is amazing, you should be very proud of what you have produced Thibaut!!

  • @ashutoshpadhi2782
    @ashutoshpadhi2782 2 месяца назад +1

    The amount of effort you put into these works is really commendable. You are a blessing to humanity.

  • @Pokojsamobojcow
    @Pokojsamobojcow 3 месяца назад +6

    Wow it must be a lot of work to obtain such amazing animations! The video is really dynamic and easy to follow, congratulations

    • @julius4858
      @julius4858 3 месяца назад

      The program is called manim, it’s from 3blue1brown

  • @AICoffeeBreak
    @AICoffeeBreak 3 месяца назад +9

    Love the clarification at 00:13, because I've also felt that the misconception is wide spread. I've heard people say: I am not using GANs anymore, I am using Generative AI. The word "Generative" is literally what G in GAN stands for. 😂😂

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад +3

      I made this intro because I couldn't stand the number of "GenAI experts" on my LinkedIn feed :(

  • @BenjaminEvans316
    @BenjaminEvans316 3 месяца назад +2

    Great video. Looking forward to your video on contrastive learning, it is my favourite subject in deep learning.
    Your videos combine great production skills (animations, colour selection, movement between frames) with in-depth understanding of complex concepts.

  • @muhammed2174
    @muhammed2174 Месяц назад

    You're a master at your craft, it is a testament to your studies!

  • @EkalabyaGhosh-g3i
    @EkalabyaGhosh-g3i 3 месяца назад +3

    This channel needs more subscribers

  • @babatundeonabajo
    @babatundeonabajo Месяц назад

    This is a really good video, and the animations are top-notch. I feel this video is good not just for those learning about AI but also those learning Statistics.

  • @Bwaaz
    @Bwaaz 3 месяца назад +1

    Amazing quality, hope the channel takes off ! Great use of manim

  • @Tunadorable
    @Tunadorable 2 месяца назад +6

    3blue1brown specifically for AI models??? sign me up!!! I'll fs be linking to you in my own vids whenever relevant, this was great

  • @ahmedoumar3741
    @ahmedoumar3741 13 дней назад

    Your videos and explanations are really excellent, please keep doing this.

  • @BeeepBo0op
    @BeeepBo0op 3 месяца назад

    Thank you for finally making me understand the reparametrization trick!! It was thrown at me several times during a DRL class I took last year and I never really understood what we did. This made it much much clearer, thank you! Also: great video overall!

  • @kacperzaleski8125
    @kacperzaleski8125 3 месяца назад +2

    cant wait for the next video! this was great!!!!!

  • @mostafasayahkarajy508
    @mostafasayahkarajy508 3 месяца назад

    Thank you very much for providing and sharing the lecture. Excelent explanation and so a high-quality video!

  • @MutigerBriefkasten
    @MutigerBriefkasten 3 месяца назад +2

    Thanky you again for the great content and the amazing animations 🎉💪👍 keep going, hopefully your Channel explode with more subscribers .. i will recommend it for sure to other people

  • @abelhutten4532
    @abelhutten4532 3 месяца назад

    Great visualizations, and good explanation! Congratulations and thanks for the nice video :)

  • @notu483
    @notu483 3 месяца назад +1

    12:40 we scale by standard deviation not variance

  • @shangliu6285
    @shangliu6285 3 месяца назад

    perfect video, its easy to understand the VAE, subscribed!

  • @carlaconti1075
    @carlaconti1075 2 месяца назад

    Thanks to you everything is clear now, thank you Deepia

  • @JONK4635
    @JONK4635 3 месяца назад

    Really amazing content, thank you for spreading knowledge! Thanks a lot :)

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      Thanks for the comment, keeps me motivated :)

  • @lucianovidal8721
    @lucianovidal8721 2 месяца назад

    Great content! It was a really good explanation

  • @cs-cs4mj
    @cs-cs4mj Месяц назад

    hey, so well explained thanks for the video!! really nailed those animations as well, would be cool to make a video on adam/rmsprop as well, i have a hard time properly understanding why they work. anyway much love to you my friend

  • @deepniba
    @deepniba 27 дней назад

    Thank you , still looking for VAE variants videos

  • @woowooNeedsFaith
    @woowooNeedsFaith 12 дней назад

    15:49 - What is convex interpolation?

    • @Deepia-ls2fo
      @Deepia-ls2fo  11 дней назад

      Basically a linear interpolation between two points, with "t" in front of one of the points, and "(1-t)" in front of the other. The set of all these points is convex, hence "convex interpolation" :)

  • @essentiallifter
    @essentiallifter 21 день назад

    just a tip can you make it super clear that the reason we sample in the middle is to produce a nice continuous latent space where the different dimensions encode different meaning

  • @HenrikVendelbo
    @HenrikVendelbo 3 месяца назад

    Thanks

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      Ho my, thank you so much !

    • @HenrikVendelbo
      @HenrikVendelbo 3 месяца назад

      I find math speak very hard to grok. I was always good at math, but always got turned off by the navel gazing and geekery. You do a great job keeping it engaging without assuming that I am a math geek

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад +1

      @@HenrikVendelbo Yeah sometimes it do be like that in math classes.
      I think it's important to look at equations when they tell us something about the models, but computational tricks or complex equations are not that interesting.

  • @abdelmananabdelrahman4099
    @abdelmananabdelrahman4099 3 месяца назад

    Great video 🎉. I've never had such a great explanation of VAE. Waiting for VQVAE.....

  • @chriskamaris1372
    @chriskamaris1372 2 месяца назад +2

    Furthermore, in 11:39 and 12:39 you are referencing σ as variance. But is it σ the standard deviation and σ^2 the variance? (Nevertheless, the video is perfect. Excellent work!)

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад

      Thanks, indeed there might be some mistakes !

  • @drannoc9812
    @drannoc9812 3 месяца назад

    Amazing content :D
    I hope you'll do your next videos on VQ VAE and VQ VAE 2, I enjoyed so much reading those papers !

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      Thanks, I really gotta take another look to the paper

  • @chcyzh
    @chcyzh 28 дней назад

    Thank you very much! It's pretty clear

  • @vitorfranca80
    @vitorfranca80 3 месяца назад

    Incredible explanation!! Thank you for sharing your knowledge! 😁😁

  • @KennethTrinh-cm6cp
    @KennethTrinh-cm6cp 3 месяца назад

    thanks for the wonderful animations and explanation

  • @flueepwrien6587
    @flueepwrien6587 3 месяца назад

    Finally I understand this concept.

  • @authenticallysuperficial9874
    @authenticallysuperficial9874 25 дней назад +1

    Audio comes out from under water at 2:09 btw

    • @Deepia-ls2fo
      @Deepia-ls2fo  25 дней назад

      Thank you, I had some issues with copyrighted music which led to RUclips removing it but also degrading the audio...

  • @nihilsson
    @nihilsson 2 месяца назад

    Great vid! Commenting for algorithmical reasons

  • @samuelschonenberger
    @samuelschonenberger 27 дней назад

    Watching this while my first VAE is training

  • @griterjaden
    @griterjaden 2 месяца назад

    Wowowowowowow 🎉🎉🎉 amazing video for VAE. Pls ~ make more videos

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад

      @@griterjaden Thanks, I'm on it :)

  • @Jay_Tau
    @Jay_Tau 3 месяца назад

    This is excellent. Thank you!

  • @markbuckler4793
    @markbuckler4793 3 месяца назад

    Excellent video, I subscribed because of it :)

  • @rishidixit7939
    @rishidixit7939 Месяц назад

    At 7:45 why is the assumption for p(z) as a Normal Distribution important ? Without that are further calculations not possible ?
    At 8:01 why is the posterior assumed to be Gaussian ?

    • @Deepia-ls2fo
      @Deepia-ls2fo  Месяц назад

      @@rishidixit7939 Hi again, indeed further calculations are intractable without assuming both the prior and the posterior to be Gaussian.
      Some other research works have replaced these assumptions by other well known distributions such as mixtures of Gaussians, which results in another training objective.

  • @neerajsingh-xf3rp
    @neerajsingh-xf3rp 11 дней назад

    0:23 does it create data from scratch?

    • @Deepia-ls2fo
      @Deepia-ls2fo  11 дней назад

      Yep, basically modern image generation techniques (diffusion models / flow matching) create new data starting from pure noise !

    • @neerajsingh-xf3rp
      @neerajsingh-xf3rp 11 дней назад

      @@Deepia-ls2fo does it learn from existing data ? , if yes how does it generate data from scratch , denoising involves learning the state and adding some randomness in that state only 🤔

  • @authenticallysuperficial9874
    @authenticallysuperficial9874 25 дней назад +1

    Thanks!

  • @michael91703
    @michael91703 3 месяца назад

    Is this manim?!!! Nice work dude!

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      It is indeed Manim, thank you !

  • @hannes7218
    @hannes7218 Месяц назад

    good stuff! keep it going

  • @English-bh1ng
    @English-bh1ng 3 месяца назад

    It is the best VAE visualization.

  • @Bikameral
    @Bikameral 3 месяца назад

    Great content ! What software are u using to animate ?

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      Thanks ! For most animations I use Manim, a python module originally made by Grant Sanderson from 3blue1brown.

    • @Bikameral
      @Bikameral 3 месяца назад

      @@Deepia-ls2fo thank you

  • @guilhermegomes4517
    @guilhermegomes4517 3 месяца назад

    Great Video!

  • @user-wr4yl7tx3w
    @user-wr4yl7tx3w 17 дней назад

    Great content

  • @Chadpritai
    @Chadpritai 3 месяца назад

    Next video on diffusion models please , thanks in advance ❤

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад +1

      It's on the to-do list but the next 3 videos will be about self-supervised learning !

  • @i2c_jason
    @i2c_jason 3 месяца назад

    Is there a statistical property or proof that might show a graphRAG "transfer function" to be the same as a VAE or maybe a CVAE? Perhaps in terms of entropy? It would be interesting to make two identical systems, one using a VAE and one using graphRAG, and see if they can match up statistically. I can't shake the idea that software 3.0 might be the more sound approach for developing new GenAI tools vs software 2.0.

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад +1

      Hi Jason ! Unfortunately I know close to nothing about RAG so I have no idea if what you describe might be feasible. I here about RAG everywhere these days, I should get up to date on that.

    • @i2c_jason
      @i2c_jason 3 месяца назад

      @@Deepia-ls2fo I'd love to hear your take on it if you ever do a deep dive.

  • @EigenA
    @EigenA 27 дней назад

    Great video. What is your educational background?

    • @Deepia-ls2fo
      @Deepia-ls2fo  27 дней назад +1

      Thanks ! Bachelor in math, bachelor in computer science, master in AI/ML, currently doing a PhD in applied maths and deep learning

    • @EigenA
      @EigenA 27 дней назад

      @ legendary. Good luck on the PhD! I’m 3rd year EE PhD student, you have phenomenal content. Looking forward to watching your channel grow.

  • @notu483
    @notu483 3 месяца назад

    Thanks for the video ❤😊

  • @3B1bIQ
    @3B1bIQ 2 месяца назад +2

    🤍Please, can you create a course to learn the manim library from scratch to professionalism, because I need it very much? Please reply ❤😊

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад +2

      Thanks for your comment, I would love to but I have many other topics I want to talk about first, and not much time on my hand! There are very good ressources on RUclips though, if you want to start to learn Manim. :)

    • @3B1bIQ
      @3B1bIQ 2 месяца назад +1

      @@Deepia-ls2fo Thank you, but I hope that you have enough time to create a course to learn manim, even if there is one video every week, and this will contribute to increasing the number of your views more because your explanation is very beautiful and clear, and I can understand it easily even though I am an Arab 🤍☺️

  • @maths.visualization
    @maths.visualization 4 дня назад +1

    Can you share video code?

  • @sillasrocha9623
    @sillasrocha9623 2 месяца назад

    Hey, could you make a video talking about swav in unsupervised learning?

  • @cupatelj52
    @cupatelj52 2 месяца назад

    great content bro.

  • @prabaldutta1935
    @prabaldutta1935 2 месяца назад

    Amazing Graphics and explanation. I have one question - if we use MNIST dataset (like what is shown in the video) does it mean that the mu and sigma are vectors of dimension 10x1? What if we use a dataset where the number of different classes are unknown? What will be the dimension of mu and sigma in that case?

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад +1

      Thank you, the latent dimension is not directly related to the number of classes in your dataset.
      In fact a very good encoder could very well classify perfectly the 10 classes on a single dimension, but it makes things way harder to reconstruct for the decoder.
      As you mention in most datasets we don't even know the number of classes or the number of relevant features, so we just take ad hoc latent dimensions (16, 32) and see if it's enough for the encoder to produce a useful representation, and for the decoder to reconstruct correctly.

    • @prabaldutta1935
      @prabaldutta1935 2 месяца назад

      @@Deepia-ls2fo Thanks a lot for your response. Can't wait for your next video.

  • @syc52
    @syc52 2 месяца назад

    Could you please make a video talking about why diffusion model, GAN, and VQVAE can make the image sharper

  • @awsaf49
    @awsaf49 2 месяца назад

    Hey, nice video!

  • @aregpetrosyan465
    @aregpetrosyan465 2 месяца назад

    This question came to my mind: What would happen if we ignored the encoder part and tried to train only the decoder? For example, by sampling from a standard Gaussian vector and attempting to reconstruct a digit. I don't really understand the purpose of the encoder.

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад

      If you don't condition at all the latent space from which you are sampling, I'm not sure the model will be able to learn anything.
      Here the encoder explicitly approximate the posterior distribution in order for us to then sample from the distribution of images.
      This is all a theoretical interpretation of course, but learning to reconstruct any digit from pure unconditioned noise seems a bit hard!
      Diffusion models kind of do it (in image space), but this usually takes a lot of steps.
      Anyway, the experiment you describe would be very easy to implement, if you want to try it out. :D

  • @rrttttj
    @rrttttj 3 месяца назад

    Great video! However, I am slightly confused: For your loss function you are subtracting KL divergence rather than adding it. Wouldn't you want to add it to penalize the difference between the latent distribution and the standard normal distribution? At least, in all implementations I have seen they add KL divergence rather than subtract it.
    Edit: I understand my mistake now!

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      Hi ! Thanks for the comment, I'm afraid I might have flipped a sign at one point.
      When you derive the ELBO (which you then maximize via training), there is a minus sign appearing in front of the KL. But in practice you minimize the opposite of this quantity, which is equivalent to minimizing the L2 plus the KL.
      I hope it's not too confusing. :)

    • @rrttttj
      @rrttttj 3 месяца назад

      @@Deepia-ls2fo Oooooh I understand, so ELBO is the quantity that should be maximized, and you were denoting the ELBO quantity with L(x), not the loss itself. I understand now, thanks!

  • @NguyenAn-kf9ho
    @NguyenAn-kf9ho 2 месяца назад

    is there we have videos with the same approach, for Reinforcement Learning :D ???? !

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад

      Hi, unfortunately I don't know anything about Reinforcement Learning, so I don't think I'll be able to make videos about that any time soon. However, I believe Steve Brunton has very good videos on the topic :)

  • @shashankjha8454
    @shashankjha8454 Месяц назад

    do u use manim for animations ?

  • @EdeYOlorDSZs
    @EdeYOlorDSZs Месяц назад

    top tier video!

  • @미비된동작-p4g
    @미비된동작-p4g 3 месяца назад

    13:05 That’s so funny VAE and Adam both are proposed by same person, Kingma..

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад +1

      He's quite the man, also co-author on some key diffusion models paper :)

  • @asteriskman
    @asteriskman 3 месяца назад

    'now that we've got the basics down' ... lol yea ok, professor.

  • @zyansheep
    @zyansheep 3 месяца назад

    Now if only the latent space could be a variable size, and be discrete, then maybe we could do effective ai lossy/lossless compression 🤔

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      Hi, I don't know about variable dimension latent space, but discrete sure sounds like VQ-VAE :)

  • @arno7198
    @arno7198 3 месяца назад

    DeepIA absolutely killed it with this video on Variational Autoencoders. As a government official, medical doctor, and law PhD, it's not often I come across something that genuinely teaches me something new. But this video? Wow.
    The way Variational Autoencoders map data to a latent distribution instead of a fixed point, and the balance between reconstruction loss and Kullback-Leibler divergence, was explained so clearly that I picked it up right away.
    Whether I'm shaping policies, treating patients, or analyzing legal cases, this video added value in ways I didn’t expect. Props to DeepIA for delivering content that even someone as busy (and brilliant) as me can appreciate!
    And let’s not forget the genius behind it all. Honestly, the mind that creates content like this is nothing short of extraordinary. I don’t say this lightly, but DeepIA might just be the most insightful, brilliant, and generous creator on RUclips. The precision, the depth, the clarity-it’s rare to find someone who can not only understand such complex topics but also make them accessible to mere mortals like us. It’s an honor to witness this level of mastery. Truly, we’re not worthy.

  • @أمديعماد
    @أمديعماد 3 месяца назад

    You need that music I dont remember it

  • @fedorzhdanov6085
    @fedorzhdanov6085 3 месяца назад +1

    There are so many trashy channels with AI generated nonsense, while some channels (like this) has clear explanation and just few views. I think RUclips should add some "peer-review" feature, and while there is no such tool I encourage support such good channels with likes and comments and hit dislikes for useless AI "blah-blah" channels.
    I'm not against AI as helper tool (like script writing/voice generation), but if there is no fact checks from authors, that make it garbage and the platform doesn't have proper garbage collector yet.

  • @alexm9606
    @alexm9606 2 месяца назад

    Comment to up this channel

  • @MursaleenFayyaz-lr7hu
    @MursaleenFayyaz-lr7hu 2 месяца назад

    Please create videos on Auto Regressive Models, Particularly RNN, LSTM, PixelCNN as soos as you can. I have a mid exam in third week of October which will cover these topics.

  • @duncanwarukira4348
    @duncanwarukira4348 3 месяца назад

    Is this voice AI

    • @Deepia-ls2fo
      @Deepia-ls2fo  3 месяца назад

      Yes I cloned my voice using a text to speech service called elevenlabs

  • @sheevys
    @sheevys 2 месяца назад

    The auto-generated voice-over is super annoying. Any chance a real human can narrate it?

    • @Deepia-ls2fo
      @Deepia-ls2fo  2 месяца назад +2

      No to be honest that would take way too much time on my side, so it's probably never going to happen. Hopefully text to speech services get better over time!

  • @timewasting7574
    @timewasting7574 Месяц назад

    3:25 - 6:20 is so distracting. Just assume your audience knows these. No need to conform your target group to general public. Just assume senior-year undergraduate please.

  • @ruslanbip7333
    @ruslanbip7333 2 месяца назад

    Thanks