Jakub Tomczak @jmtomczak Twitter profile | Pikagi

Pikagi

Jakub Tomczak

@jmtomczak

8,611

Followers

985

Following

229

Media

1,725

Statuses

Group Leader & Principal Scientist | @TUeindhoven | @natinlab1 | Before: @Qualcomm @VUamsterdam @UvA_Amsterdam ( @wellingmax ) | Opinions my own

Amsterdam, the Netherlands

https://t.co/VPda9AjCyc

Joined October 2016

Don't wanna be here? Send us removal request.

Pinned Tweet

@jmtomczak

Jakub Tomczak

9 days

🎊 It has arrived 🎊, the 2nd edition of my "Deep Generative Modeling" book. It has 100 new pages, 3 new chapters (incl. #LLMs ) and new sections. It covers all deep generative models that constitute the core of all #GenerativeAI techs! Check it out: 💻

Tweet media one

10

82

553

Last Seen Profiles

@JohnMorello7

@rosemberglayla1

@HK_29_OV

@kaga_rin0321

@alwolfk55

@rosy_kbee

@ravuina

@mujeres_psuv

@nanaGU_AI

@taimatsu_18

@MoistChi

@STEENMS2

@sailorlewd11

@Grady97546986

@shisou_

@e1spo

@sharif_dar34865

@samantoun

@JJysq8vvfj11645

@RKNPARS

@edsheeran

@Q4920384051241

@jrBarbozaArnold

@Dreamzy23

@OsenPok17426

@CanteraCDDB

@M07200527

@alendadoheroi

@caixukun168667

@pyramiDAO

@lahtisymphony

@Olive_EvansNZ

@BinorRaja

@ps

@stw46

@waffohs

@jmtomczak

Jakub Tomczak

3 years

I am excited to announce that my book, "Deep Generative Modeling", is available online and in print ( @SpringerNature ): Code used in the book is freely available online: (1/4)

Tweet media one

34

336

2K

@jmtomczak

Jakub Tomczak

4 years

New year, and new challenges! I've decided to do a series of blog posts on deep generative modeling. The first post is out: I hope you'll enjoy it! The next one in ~2 weeks 🤓

Tweet media one

Tweet media two

Tweet media three

15

129

694

@jmtomczak

Jakub Tomczak

4 years

The fourth blog post on deep generative modeling! This time about one of my favorite topics: Variational Auto-Encoders ▶️◀️. Proposed concurrently by @dpkingma @wellingmax and @DaniloJRezende @shakir_za D. Wierstra. 👇 Post: Code:

Tweet media one

Tweet media two

Tweet media three

8

86

413

@jmtomczak

Jakub Tomczak

5 months

I am extremely happy to announce that yesterday, after ~12mo of hard work, I handed in the second edition of my book "Deep Generative Modeling" () to the publisher. There are ~100 pages of new content! ✨ 💡I will keep you updated!!💡

Tweet card media

Deep Generative Modeling

link.springer.com

9

62

404

@jmtomczak

Jakub Tomczak

3 years

A new blog post is out! This time about Diffusion-based Deep Generative Modeling! Read about this fascinating research line and check out the code: Post 📜: Code 💻:

Tweet media one

Tweet media two

Tweet media three

5

76

396

@jmtomczak

Jakub Tomczak

3 years

The 7th post in the series on deep generative modeling. This time about priors ⏺️ in VAEs ▶️◀️, a very important topic that is often neglected. Join me in the journey on standard Gaussians, VampPriors, flows and GTMs. 👇 Post:

Tweet media one

Tweet media two

Tweet media three

8

58

367

@jmtomczak

Jakub Tomczak

3 years

Yesterday, in our deep learning course () at @VUamsterdam , we had an amazing guest lecture on "Geometric Deep Learning" provided by @TacoCohen . More on this topic could be found at ( @mmbronstein @joanbruna @TacoCohen @PetarV_93 )

Tweet media one

4

50

347

@jmtomczak

Jakub Tomczak

2 years

Can we efficiently learn one model that generates and classifies (i.e., a joint model)? Yes, we can! In our recent paper, we proposed a new joint model that combines a diffusion model with a classifier by sharing the UNet architecture. Link : [1/5]

Tweet media one

9

49

340

@jmtomczak

Jakub Tomczak

3 years

After the summer break, the deep generative modeling series is back! 💪 This time about hierarchical latent variables models (hierarchical VAEs). Post 📜: Code 💻:

Tweet media one

Tweet media two

Tweet media three

4

58

342

@jmtomczak

Jakub Tomczak

3 years

After a long time, the 8th blog post is out! This time I discuss neural compression with deep generative modeling. Post 🤓: Code 💻:

Tweet media one

Tweet media two

Tweet media three

4

64

308

@jmtomczak

Jakub Tomczak

7 months

The end of February brings the last blog post in the short series on score-based generative models. After score matching and SDEs, it is time for flow matching! Link to the post: Link to the code:

4

58

269

@jmtomczak

Jakub Tomczak

2 years

Exactly one year ago, after writing it for 9mo (yep, it's my baby!), my book on #GenerativeAI was out! It was downloaded over 26k, which is over 70 times per day! It was an amazing year for #GeneartiveAI , and if you'd like to catch up with this topic, just get this book 🤓

@jmtomczak

Jakub Tomczak

3 years

I am excited to announce that my book, "Deep Generative Modeling", is available online and in print ( @SpringerNature ): Code used in the book is freely available online: (1/4)

Tweet media one

34

336

2K

2

29

268

@jmtomczak

Jakub Tomczak

3 years

Is there a better way to start a week than reading about deep generative modeling? This time about energy-based models! Post 📜: Code 💻:

Tweet media one

Tweet media two

3

58

258

@jmtomczak

Jakub Tomczak

9 months

Today is a new day and a new blog post on score-based generative models. How do we use score matching to learn generative models defined by a stochastic/ordinary differential equation? Check this out! Post: Code:

Tweet media one

2

54

257

@jmtomczak

Jakub Tomczak

2 years

Today we had a great e-talk by @LucaAmb for AI ppl at the @VUamsterdam ! Extremely interesting stuff on #deepgenerativemodeling , a combination of VAEs and flows. More details in their writeup:

Tweet media one

4

41

253

@jmtomczak

Jakub Tomczak

2 years

Kudos to Dr. @emiel_hoogeboom for his amazing Ph.D. project defended today! A #GenerativeAI superstar in all aspects! P.S. I have never heard so witty and lovely laudatio (given by @wellingmax ) before.

Tweet media one

8

19

250

@jmtomczak

Jakub Tomczak

4 months

I played around a bit with KANs (Kolmogorov-Arnold Nets) parameterizing a VAE on a tiny data (1797 8x8 images of handwritten digits). Here are the results (ELBOs): MLP (~0.5M): 96.8 KAN (~4M): 97.2 MLP+KAN (~2.7M): 96.7 KAN+MLP (~1.6M): 97.5 KAN (enc) and MLP (dec): 96.9

Tweet media one

8

33

225

@jmtomczak

Jakub Tomczak

6 years

The updated version of our Hyperspherical VAE is out (). We have both @TensorFlow code () and @PyTorch code as well (). More info on the blogpost: .

Tweet media one

2

70

206

@jmtomczak

Jakub Tomczak

6 months

After the score-based models, I decided to take a step back and cover the basics in a new blog post: Probabilistic modeling and Mixture Models. Additionally, I make a brief intro to ✨Probabilistic Circuits✨ Check: 📄Post: 🖥️ Code:

Tweet media one

Tweet media two

Tweet media three

3

38

206

@jmtomczak

Jakub Tomczak

9 months

New year, new blog posts! Today, I start with the first post from a series on score-based generative models. Are you curious about score matching and how to implement it? Check this out! Post: Code:

Tweet media one

3

35

192

@jmtomczak

Jakub Tomczak

4 years

My recent work on general invertible transformations for flow-based generative models with some preliminary experiments. I present how we can generalize coupling layers and reversible logic operators. Paper: Code:

Tweet media one

Tweet media two

1

33

190

@jmtomczak

Jakub Tomczak

4 years

The sixth blog post on deep generative modeling! This time about hybrid modeling for p(x,y). Mainly based on great work from @eric_nalisnick @amatsukawa @yeewhye D. Gorur & @balajiln ! 👇 Post: Code:

Tweet media one

Tweet media two

Tweet media three

1

47

182

@jmtomczak

Jakub Tomczak

4 years

The fifth blog post on deep generative modeling! This time about Integer Discrete Flows. Proposed by my friends from Amsterdam: @jornpeters @emiel_hoogeboom @vdbergrianne @wellingmax 👇 Post: Code:

Tweet media one

Tweet media two

Tweet media three

2

38

177

@jmtomczak

Jakub Tomczak

6 years

A magnificent paper on hybrid models using well-known generalized linear models and recent developments on flow-based generative models by @eric_nalisnick , @amatsukawa , @yeewhye , D. Gorur & @balajiln : . This is why generative modeling is important!

Tweet media one

Tweet media two

Tweet media three

5

55

174

@jmtomczak

Jakub Tomczak

6 years

A beautiful paper by Bin Dai & David Wipf "Diagnosing and Enhancing VAE Models" ()! One of best appendixes I've ever read. Must read for anyone interested in VAEs. Thank you @tolstikhini for pointing out this paper!

Tweet media one

Tweet media two

2

36

176

@jmtomczak

Jakub Tomczak

4 years

As promised, today a new blog post is out! This time, we start with the first class of deep generative models in the series: Deep Autoregressive Models. Additionally, I provide a code that you could play around with! Post: Code:

Tweet media one

Tweet media two

Tweet media three

3

41

170

@jmtomczak

Jakub Tomczak

2 years

Thank you @NeurIPSConf for a great birthday present: 2 out of 2 papers accepted! Congrats to the coauthors: @a_kzna @wellingmax @mrpejcio @tomasztrzcinsk1 . @a_kzna and @mrpejcio , many thanks for your amazing work!

14

6

159

@jmtomczak

Jakub Tomczak

3 years

A beautiful quote from Shannon. I have an impression that the AI field goes against this piece of wisdom. Companies and unis expect students to publish papers. Postdocs and asst. profs are expected to deliver >1 papers a year. We arrive at >9k submissions per conf. #time4change

Tweet media one

2

25

148

@jmtomczak

Jakub Tomczak

3 years

A very interesting special issue on probabilistic methods for deep learning (deadline October, 2021) edited by @eric_nalisnick and @dustinvtran : 👉

Tweet card media

Probabilistic Methods for Deep Learning

Special Issue in journal Entropy: Probabilistic Methods for Deep Learning

2

30

149

@jmtomczak

Jakub Tomczak

8 months

It's going to happen! What? Generative Modeling Summer School 2024 ( #GeMSS24 )! When: June 24-28, 2024 Where: Eindhoven, the Netherlands ( @TUeindhoven , @Brainport_int ) Info: Application deadline: March 31, 2024 Don't wait, apply!

Tweet media one

8

40

147

@jmtomczak

Jakub Tomczak

2 years

I decided to refresh my website () and... start a new (short) series of blog posts! Today the kickoff post: "Trouble in paradise: Does it make sense to train latent variable models with variational inference?" Post:

Tweet media one

5

19

148

@jmtomczak

Jakub Tomczak

4 years

Today the next blog post is out! This time, flow-based models are discussed, with RealNVP as an example. Blog post: Blog: Code:

Tweet media one

Tweet media two

Tweet media three

1

35

146

@jmtomczak

Jakub Tomczak

10 months

Today we start @NeurIPSConf 2023 in amazing New Orleans! Morever, I am excited to announce that I will be a program chair of @NeurIPSConf 2024 together w/ Cheng Zhang & @ulrichpaquet , and @DaniCMBelg as the general chair. Great days ahead 💯

Tweet media one

10

4

147

@jmtomczak

Jakub Tomczak

2 years

OK, so if anyone wants to delve into diffusion-based models, there is a cool paper about it by @calvinyluo : And (shameful self-promotion) you can always first take a look at my book #deepgenerativemodeling :

2

32

141

@jmtomczak

Jakub Tomczak

4 years

1/n The NeurIPS 1st round of reviewing is coming. Again, I see more and more comments on Twitter like "please don't use another 'theorem' if it makes no sense", unclear theory doesn't make you paper better" or "English is so poor, why did you submit the paper".

2

12

145

@jmtomczak

Jakub Tomczak

3 years

The paper "Consistency Regularization for Variational Auto-Encoders" by @_sam_sinha_ and @adjiboussodieng is by far the neatest paper I've read in 2021! So clean and smart idea, so amazing execution! A highly recommended read:

Tweet card media

Consistency Regularization for Variational Auto-Encoders

Variational auto-encoders (VAEs) are a powerful approach to unsupervised learning. They enable scalable approximate posterior inference in latent-variable models using variational inference (VI)....

8

41

138

@jmtomczak

Jakub Tomczak

3 years

Since today, I am an @ELLISforEurope member! Thank you @zeynepakata and @tameem_adel for your endorsement.

Tweet media one

10

3

136

@jmtomczak

Jakub Tomczak

2 years

I heard from people at @RealAAAI 2023 that they have very good books for sale😉

Tweet media one

2

7

127

@jmtomczak

Jakub Tomczak

2 years

When you wake up to read this email from ICLR 😍 Congrats and thanks for the collaboration @davidwromero @davidmknigge @_albertgu @egavves @erikjbekkers M. Hoogendorn & JJ Sonke! 🙏 Paper:

Tweet media one

10

2

128

@jmtomczak

Jakub Tomczak

1 year

I am extremely excited to announce an opening (Ph.D. candidate) in my group at @TUeindhoven : Please apply here: Join us in the fantastic journey of developing #GenerativeAI !

Tweet card media

This job is unavailable

The TU/e is constantly looking for scientific and non-scientific staff further its ambitions. View here our current vacancies.

1

41

129

@jmtomczak

Jakub Tomczak

3 months

We have kicked off! The best summer school on #GenerativeAI : ✨Generative Modeling Summer School 2024✨! We turn @TUeindhoven into the world's capitol of #GenAI . Amazing lineup: @jesfrellsen , @pamattei , @dpkingma , @liyzhen2 , @helibenhamu , @StephanMandt , @ewa_szczurek , @vlamen 🤩

Tweet media one

Tweet media two

5

20

126

@jmtomczak

Jakub Tomczak

3 months

The evaluations of my course came in. Here are some cherry-picked opinions: 🍒 "Jakub is an amazing teacher, which made the course very fun" "Jakub is passionate about the field, and it makes the lectures very enjoyable" "The lecturer was very knowledgeable and entertaining" 🫳🎤

Tweet media one

8

1

124

@jmtomczak

Jakub Tomczak

6 years

A neat idea on utilizing the accept/reject sampling as the prior in the VAE by @MatthiasSBauer and @AndriyMnih "Resampled Priors for Variational Autoencoders".

Tweet media one

3

27

123

@jmtomczak

Jakub Tomczak

5 years

I'm super excited about joining @CIGroupVU at @VUamsterdam as an assistant professor of AI. Looking forward to working with @gusz_e , @FlorisHengst , @KarineSMiras , @GongjinLan , @eliseofe , @ali_elhassouni , @matteo_phd , @AleZonta , @EmilevanKrieken , @jgKlein Mark, Eoin and others!

12

4

122

@jmtomczak

Jakub Tomczak

2 years

A wonderful day with Dr. @MaxIlse ! Today Maxi had his defense and I am a proud co-promotor (together with @wellingmax ). Congratulations, Maxi!

Tweet media one

13

1

122

@jmtomczak

Jakub Tomczak

3 years

Once the whole NeurIPS dust is down, it's time to return to the series on deep generative modeling! This time about GANs, a class of models that made a lot of fuzz in the world of DGMs. Post 📜: Code 💻:

Tweet media one

Tweet media two

Tweet media three

2

26

121

@jmtomczak

Jakub Tomczak

4 months

The deadline for @NeurIPSConf 2024 has passed, but it doesn't mean we don't work hard to have another fantastic conference! Right now, we are in need for REVIEWERS. If you know someone suitable for this role, please fill in this form:

Tweet card media

Reviewer suggestion/self-nomination for NeurIPS 2024

This form serves a single purpose: Providing a name and

docs.google.com

6

32

120

@jmtomczak

Jakub Tomczak

5 years

Our final version of the @NeurIPSConf paper on Combinatorial Bayesian Optimization is available: + CODE: . An amazing work done by @ocy0224 and a great collaboration with @egavves & @wellingmax . Joint project: @quvalab - @Qualcomm_Tech

Tweet media one

Tweet media two

Tweet media three

Tweet media four

0

28

111

@jmtomczak

Jakub Tomczak

2 years

Today is the second blog post in the new series. This time: "Are we dil(ff)usional about the ELBO?!". What are the consequences of using Gaussian diffusion in latent variable models? Blog: Post: #deepgenerativemodeling #diffusion

Tweet media one

2

19

112

@jmtomczak

Jakub Tomczak

6 years

Two wonderful papers from @oxcsml ( @yeewhye group): "Hierarchical Representations with Poincaré Variational Auto-Encoders" () and "Probabilistic symmetry and invariant neural networks" ()!

Tweet media one

Tweet media two

1

26

109

@jmtomczak

Jakub Tomczak

5 years

Our paper "Conditional Channel Gated Networks for Task-Aware Continual Learning" got accepted to CVPR 2020! An amazing work done by @davideabati with @Qualcomm_Tech ( @TiRune , @BabakEht , me) and @UNIMORE_univ (S. Calderara & R. Cucchiara). Working with Davide was a pure pleasure!

Tweet media one

Tweet media two

Tweet media three

5

26

108

@jmtomczak

Jakub Tomczak

10 months

Day 5 of Neurips: Greetings from the one and only @wellingmax to all the folks out there!

Tweet media one

2

1

109

@jmtomczak

Jakub Tomczak

2 years

Energy-based models (in the form of Gaussian Restricted Boltzmann Machines) strike again! A very interesting paper by @lrjconan @skornblith @mengyer @fleet_dj & @geoffreyhinton : w/ @PyTorch code: #deepgenerativemodeling

Tweet media one

2

12

106

@jmtomczak

Jakub Tomczak

1 month

📢 Keynote speakers for @NeurIPSConf 2024 have now been confirmed! 📢 We are ✨ excited ✨ to welcome @HochreiterSepp , @DanicaKragic , Lidong Zhou, @AlisonGopnik , @RosalindPicard , @ArnaudDoucet1 , and @drfeifei as speakers this year!

0

16

101

@jmtomczak

Jakub Tomczak

2 years

Together with @jesfrellsen & @pamattei , we organize Generative Modeling Summer School #GeMSS23 ! Please check out our webpage and apply! Deadline: April 10, 2023 Where: Copenhagen (Denmark) When: June 26-30, 2023 Lecturers: see below👇

Tweet media one

3

25

94

@jmtomczak

Jakub Tomczak

3 years

If you are looking for your niche in AI, here are five topics to avoid 😉 Very nice to see deep generative models (diffusion models) to be listed out.

Google Research: Themes from 2021 and Beyond

Posted by Jeff Dean, Senior Fellow and SVP of Google Research, on behalf of the entire Google Research community Over the last several decades, I'v...

research.google

3

12

93

@jmtomczak

Jakub Tomczak

8 months

Our paper on Wavelet Nets got accepted to @TmlrOrg @TmlrPub ! I am extremely happy about it for several reasons: 1) we never gave up, 2) it is a nice idea imho, 3) it is one of 17 papers published (crazy!) by @davidwromero during his PhD! Congrats David, Erik and Mark 👏

@TmlrPub

Accepted papers at TMLR

8 months

Wavelet Networks: Scale-Translation Equivariant Learning From Raw Time-Series David W. Romero, Erik J Bekkers, Jakub M. Tomczak, Mark Hoogendoorn. Action editor: Jeremias Sulam. #wavelet_networks #wavelet #transforms

0

10

41

2

6

90

@jmtomczak

Jakub Tomczak

1 year

I am very excited to announce a (junior) postdoc position in my Generative AI group at @TUeindhoven ! More details: Deadline: May 14, 2023

Tweet card media

This job is unavailable

The TU/e is constantly looking for scientific and non-scientific staff further its ambitions. View here our current vacancies.

1

25

91

@jmtomczak

Jakub Tomczak

7 years

Train neural nets by using Adam with normalized gradients! Really nice paper:

1

30

87

@jmtomczak

Jakub Tomczak

1 year

I am extremely excited about our ( @Adam_Izdebski , @TomczakWeglarz , @ewa_szczurek ) new paper on Joint Transformers for De Novo drug design! You can check out the paper here: 1/n

1

10

86

@jmtomczak

Jakub Tomczak

3 years

Finally, the paper on self-supervised VAEs is published in @Entropy_MDPI 🔥 A great work w/ @JohnGatop on incorporating deterministic and non-learnable transformations (e.g., downscaling, edge detection) in VAEs to produce sharper images. Paper 👉:

Tweet media one

Tweet media two

Tweet media three

2

13

84

@jmtomczak

Jakub Tomczak

8 months

I feel delighted about "Attention-based Multi-instance Mixed Models" being accepted to @aistats_conf as an oral! It is our attention-based MIL w/ @MaxIlse on steroids: with pre-trained embeddings, variational inference and SOTA incl. histopathology! Paper:

Tweet card media

Mixed Models with Multiple Instance Learning

Predicting patient features from single-cell data can help identify cellular states implicated in health and disease. Linear models and average cell type expressions are typically favored for this...

6

9

84

@jmtomczak

Jakub Tomczak

5 months

Our paper (Mixed Models for Multiple Instance Learning) got an ORAL and OUTSTANDING STUDENT PAPER at AISTATS 2024 🎉💯🎉🚀 📃: 💻: Congrats to @janxengelmann @ale__palmaa @fabian_theis @fpcasale 👏 Hard work pays off 🔥

Tweet card media

GitHub - AIH-SGML/mixmil: Code for the paper: Mixed Models with Multiple Instance Learning

Code for the paper: Mixed Models with Multiple Instance Learning - AIH-SGML/mixmil

@janxengelmann

Jan Engelmann

5 months

Mixed Models with Multiple Instance Learning (MixMIL) received an Oral & Outstanding Student Paper award at @aistats_conf last week! 🏆 MixMIL enables accurate & interpretable patient label prediction from single-cell data by adding attention to GLMMs. #singlecell #MachineLearning

Tweet media one

Tweet media two

Tweet media three

2

12

53

2

6

82

@jmtomczak

Jakub Tomczak

1 year

The Chinese translation of my book on #GenerativeAI "Deep generative modeling" is out! It looks gooood... 🤩 Link:

Tweet media one

0

3

82

@jmtomczak

Jakub Tomczak

2 years

The NeurIPS dust is down and now it's time to... apply to the asst. prof. of ML position at @VUamsterdam 🤓 Come to work in Amsterdam 🇳🇱, the city hub of AI 💻🤖, with many brilliant ppl. 👩‍🦱👦🏾👨‍🦲🧑🏿‍🦱 Details:

1

24

82

@jmtomczak

Jakub Tomczak

4 years

An amazing work done by @yperugachidiaz on developing invertible dense networks (i-DenseNets) with learnable concatenation, an extension of Residual Flows (Chen et al., 2019). Paper: Code coming soon 😉

Tweet media one

Tweet media two

Tweet media three

2

17

81

@jmtomczak

Jakub Tomczak

2 years

If it's raining or snowing (or both 🙈) it always feels like 🌞 at @TUeindhoven ! Especially with the new #TUeLogo !

Tweet media one

4

0

80

@jmtomczak

Jakub Tomczak

1 month

Since Feb 2024, I have received ~3300 emails regarding @NeurIPSConf 2024. I can't wait Dec 2024 🙈

3

3

78

@jmtomczak

Jakub Tomczak

4 years

I am extremely proud to share a work on Approximate Bayesian Computation for discrete data (w/ Ilze A. Auzina). Applications: learning QMR-DT Network, (toyish) binary neural nets, and Neural Architecture Search. Paper: Code:

Tweet media one

Tweet media two

3

15

76

@jmtomczak

Jakub Tomczak

2 years

The last day at the @VUamsterdam with so many amzing ppl @jo_brandstetter @davidwromero @EmilevanKrieken @ayaman00 @ThiviyanSingam @KarineSMiras @FudaVan and many others! Thanks😍 @EindhovenUni here I come!

Tweet media one

4

2

76

@jmtomczak

Jakub Tomczak

1 year

Aaaaannnnnddddd... #GeMSS23 has started! @jesfrellsen @pamattei

Tweet media one

3

7

77

@jmtomczak

Jakub Tomczak

7 years

Generating images from patches: Only MNIST so far but isn't it interesting?

Tweet media one

0

19

73

@jmtomczak

Jakub Tomczak

3 years

For everyone who missed the talk or wants to re-watch it, the slides and the recording of the talk are available online (): Slides: Video: Thank you @TIIuae for the invitation!

Tweet card media

AI Seminar Series - Jakub Tomczak

Jakub Tomczak, Assistant Professor of Artificial Intelligence in the CI GroupDate of seminar: March 8, 2022 Talk topic: Deep Generative Modeling is a key to ...

www.youtube.com

@jmtomczak

Jakub Tomczak

3 years

I am happy to give a talk about deep generative modeling at @TIIuae next week on March 8. I will discuss why DGM is a key to unlocking the potential of AI! More info here: Hope to see you there!

Tweet media one

1

10

55

3

16

74

@jmtomczak

Jakub Tomczak

7 years

Our ( @MaxIlse @wellingmax ) latest work on an attention-based deep multiple instance learning with application to images and histopathology scans is on arXiv now:

Tweet media one

Tweet media two

Tweet media three

4

25

75

@jmtomczak

Jakub Tomczak

5 years

Our new paper on domain invariance using VAEs w/ @MaxIlse @ChrLouizos & @wellingmax . Very interesting results on disentanglement & domain adaptation and a real application to medical imaging. Simply be like a DIVA!

Tweet media one

Tweet media two

Tweet media three

1

15

72

@jmtomczak

Jakub Tomczak

5 months

GenAI does not stop, so it is time for a new blog post. Since LLMs are everywhere, I decided to take a look at them, my curious readers. And a bonus: An implementation of a teenyGPT 🤖✨ 📄Post: 🖥️Code:

Tweet media one

1

18

72

@jmtomczak

Jakub Tomczak

3 months

The 2nd day of the best summer school on GenAI: Generative Modeling Summer School #GeMSS24 . After the morning session on ARMs (convs & transformers) given by yours truly, now we have a session on normalizing flows by @jesf

Tweet media one

3

7

71

@jmtomczak

Jakub Tomczak

3 years

Aaaaaannnnnnddddd... now it's accepted to ICML! Congrats @MaxIlse and Patrick Forre and thank you for an amazing collaboration! It's a great reward for perseverance. #NeverGiveUp 💪

@MaxIlse

Maximilian Ilse

4 years

3 rejections (UAI, NeurIPS, AAAI) and 6 months later, I updated the arxiv version of our paper: Selecting Data Augmentation for Simulating Interventions. We have more theory and propose a simple algo to select data augmentations for domain generalization!

2

8

98

3

5

70

@jmtomczak

Jakub Tomczak

3 years

Big thanks to the organizers of #GenU2021 for the invitation! It was a great pleasure to be in #Copenhagen and discuss deep generative models with so brilliant people! If you missed my talk, here are the slides: PDF 🖥️:

Tweet media one

Tweet media two

3

4

71

@jmtomczak

Jakub Tomczak

7 years

Compression of neural networks using L0 norm is not differentiable. Or it is? @ChrLouizos @wellingmax and @dpkingma show how to do that! A neat paper on learning sparse neural networks using L0 regularization:

Tweet media one

1

27

69

@jmtomczak

Jakub Tomczak

2 years

Reading such reviews of the #DeepGenerativeModeling book gives me extra power! Thank you, Brent! Book:

Tweet card media

Deep Generative Modeling

link.springer.com

@bmspell

Brent M. Spell

2 years

I just finished Deep Generative Modeling by @jmtomczak . Covering topics from VAEs to GANs to EBMs, it has changed the way I think about ML. This book also demonstrates the reasoning behind the pesky sqrt(2) voodoo that shows up in so many neural networks.

4

17

220

0

7

68

@jmtomczak

Jakub Tomczak

2 years

Wow! I must say it’s impressive! Good job, folks! I will take a closer look at that library, and recommend it to anyone working on autoencoders.

@CChadebec

Clément Chadebec

2 years

After 8 months of long coding nights ☕️ we finally officially release Pythae 🥳, a python library unifying generative autoencoder implementations including vaegan🥗, vqvae or RAEs. 🖥️ github repo: 👉paper:

Tweet media one

11

155

967

0

8

67

@jmtomczak

Jakub Tomczak

3 years

I am so privileged to work with such amazing PhD students like @davidwromero & @a_kzna ! Two papers (led by @davidwromero ) accepted to #ICLR2022 :

Tweet card media

FlexConv: Continuous Kernel Convolutions with Differentiable Kernel Sizes

When designing Convolutional Neural Networks (CNNs), one must select the size\break of the convolutional kernels before training. Recent works show CNNs benefit from different kernel sizes at...

3

6

69

@jmtomczak

Jakub Tomczak

7 years

"VAE with a VampPrior": new code () and new results ()!

Tweet media one

Tweet media two

Tweet media three

0

16

68

@jmtomczak

Jakub Tomczak

7 years

"An Information-Theoretic Analysis of Deep Latent-Variable Models" () by A.Alemi, @poolio , I.Fischer, J.Dillon, R.Saurous & K.Murphy is my favorite paper from the last year. A great insight into representation learning by analysing the mutual information.

Tweet media one

Tweet media two

Tweet media three

0

11

66

@jmtomczak

Jakub Tomczak

5 months

In the era of #GenerativeAI , it is time to talk about GenAI systems. What are the components of GenAI Systems (GenAISys)? Does the future of AI rely on GenAISys (e.g., agentic AI 🤖)? If you are interested, I'm happy to share my thoughts with you: ✨

Tweet media one

1

12

67

@jmtomczak

Jakub Tomczak

11 months

Great opportunities for prospective postdocs in our Data & AI cluster ()! Don't think twice, apply!

Tweet media one

0

20

65

@jmtomczak

Jakub Tomczak

4 years

I've just heard a very sad information that Prof. Conway passed away... As a ML researcher, I decided to make a tribute to him in the only way I could think of.

0

5

67

@jmtomczak

Jakub Tomczak

9 months

Congratulations to Dr. @EmilevanKrieken for defending his fantastic PhD WITH HONORS (cum laude)! Only ~5% of PhDs are distinguished with cum laude in the NL. As a co-promotor, I couldn't be more proud 🥲

Tweet media one

Tweet media two

2

1

63

@jmtomczak

Jakub Tomczak

1 year

Our paper "A-NESI: A Scalable Approximate Method for Probabilistic Neurosymbolic Inference" by @EmilevanKrieken @ThiviyanSingam , A. Ten Teije, @FrankVanHarmele and myself got accepted to @NeurIPSConf ! Congrats to my amazing colleagues! Paper:

Tweet media one

Tweet media two

Tweet media three

Tweet media four

1

9

65

@jmtomczak

Jakub Tomczak

5 years

I'm a bit late to the party, but I'm super excited about our paper on Bayesian optimization for combinatorial problems accepted to @NeurIPSConf . A great collaboration between @Qualcomm and @quvalab (ChangYong Oh, @egavves , @wellingmax ). + I was selected top 400 reviewers 😏

1

3

65

@jmtomczak

Jakub Tomczak

5 years

Are you looking for a PhD position? Do you want to work on continual learning, deep generative modeling and many more? Apply now: If you want to work with me, pick up the project No. 10! More info:

Tweet media one

3

22

64

@jmtomczak

Jakub Tomczak

1 year

I am very happy to be among TMLR Expert Reviewers🥇, i.e., reviewers for @TmlrOrg who have done exemplary work in evaluating @TmlrOrg submissions. Thank you 🙏 The full list:

0

1

63

@jmtomczak

Jakub Tomczak

2 months

Recently, I have been getting more interested in #GenerativeAISystems ( #GenAISys ). I am convinced that a systems-based perspective on #GenAI is crucial to design more reliable, robust, and responsive #GenAI . You can read my perspective here:

Tweet media one

Tweet media two

Tweet media three

1

8

63

@jmtomczak

Jakub Tomczak

1 year

We push #GenerativeAI forward! We updated our paper on joint diffusion models, and now we have new SOTAs on image classification and image synthesis in the class of joint models! Moreover, we provide the code (many thanks to @kamilrdeja ):

Tweet media one

Tweet media two

@jmtomczak

Jakub Tomczak

2 years

Can we efficiently learn one model that generates and classifies (i.e., a joint model)? Yes, we can! In our recent paper, we proposed a new joint model that combines a diffusion model with a classifier by sharing the UNet architecture. Link : [1/5]

Tweet media one

9

49

340

3

5

63

@jmtomczak

Jakub Tomczak

4 years

My brilliant MSc student @AmandaIlze has her first paper accepted! I am very happy that our paper on Approximate Bayesian Computation for discrete spaces got accepted to @Entropy_MDPI . 👇 Code: Paper:

Tweet card media

Approximate Bayesian Computation for Discrete Spaces

Many real-life processes are black-box problems, i.e., the internal workings are inaccessible or a closed-form mathematical expression of the likelihood function cannot be defined. For continuous...

3

8

60

@jmtomczak

Jakub Tomczak

6 years

This paper: (by @larsmaaloe @fraccarom @LieValentin & @OleWinther1 ) was put on arXiv two days ago and it was already discussed at @UvA_Amsterdam . It is wonderful and extremely well-written, all experiments are carefully carried out. Highly recommended!

Tweet media one

Tweet media two

Tweet media three

2

11

60

@jmtomczak

Jakub Tomczak

4 months

A simple code to play with it:

Tweet card media

GitHub - jmtomczak/vae_kan_example: A simple example of VAEs with KANs

A simple example of VAEs with KANs. Contribute to jmtomczak/vae_kan_example development by creating an account on GitHub.

1

20

58

@jmtomczak

Jakub Tomczak

4 years

A new exciting work w/ amazing @JohnGatop and M. Stol on generating crisp images using VAEs. The trick is to use image downscaling as a variational posterior and then super-resolution. @CIGroupVU + @BrainCreators Paper: Code:

Tweet media one

Tweet media two

1

11

58

@jmtomczak

Jakub Tomczak

7 months

Don't forget about applying to Generative Modeling Summer School on 24 - 28 Jun 2024 (Eindhoven, NL):

3

11

56