Picture

Sander Dieleman @sedielem

Research Scientist at Google DeepMind. I tweet about deep learning (research + software), music, generative models (personal account). sander.ai London, England Joined December 2014
  • Tweets
    2K
  • Followers
    50K
  • Following
    2K
  • Likes
    10K
Sander Dieleman retweeted
cvgworkshop Profile Picture

Controllable Video Generation Workshop @ ICML2024 @cvgworkshop

2 weeks ago

We are pleased to announce the first *controllable video generation* workshop at @icmlconf 2024! 📽️📽️📽️ We welcome submissions that explore video generation via different modes of control (e.g. text, pose, action). Deadline: 31st May AOE Website: sites.google.com/corp/view/cvgi…

sedielem Profile Picture

Sander Dieleman @sedielem

3 weeks ago

Text-to-music is having a moment👀 The team behind Udio are some of the brightest and most goal-driven people I've had the pleasure to work with, before they went on to found Uncharted Labs. Amazing to see the fruits of their labour out in the open!

udiomusic Profile Picture

udio @udiomusic

3 weeks ago

Text-to-music is having a moment👀 The team behind Udio are some of the brightest and most goal-driven people I've had the pleasure to work with, before they went on to found Uncharted Labs. Amazing to see the fruits of their labour out in the open!

sedielem Profile Picture

Sander Dieleman @sedielem

4 weeks ago

10 years ago to the day, I published my first ML-related blog post: sander.ai/posts/ My blogging has been very sporadic over the years, but sharing what I've learnt has been very rewarding, and probably a pretty good career move as well😁 I highly recommend it!

sedielem Profile Picture

Sander Dieleman @sedielem

a month ago

This blog post is an amazing exposition and analysis of consistency models, and how they relate to diffusion models, leading to several suggested improvements to the training procedure that look very promising. Definitely worth a read!

ZhengyangGeng Profile Picture

Zhengyang Geng @ZhengyangGeng

a month ago

This blog post is an amazing exposition and analysis of consistency models, and how they relate to diffusion models, leading to several suggested improvements to the training procedure that look very promising. Definitely worth a read!

Sander Dieleman retweeted
EEMLcommunity Profile Picture

EEML @EEMLcommunity

a month ago

DEADLINE March 29: prepare and submit your application for EEML 2024, Novi Sad, Serbia eeml.eu 🇷🇸. Topics: Basics of ML, Multimodal learning, NLP, Advanced DL architectures, Generative models, AI for Science. Check our stellar speakers! Scholarships available! 🎉

Sander Dieleman retweeted
jaakkolehtinen Profile Picture

Jaakko Lehtinen @jaakkolehtinen

a month ago

EDM2: Analyzing and Improving the Training Dynamics of Diffusion Models (CVPR 2024) Or, getting architectural details right makes diffusion models better and yields new ImageNet SOTA Paper: arxiv.org/abs/2312.02696 Blog: developer.nvidia.com/blog/rethinkin… Code: github.com/NVlabs/edm2 1/6

Sander Dieleman retweeted
AlbyHojel Profile Picture

Alberto Hojel @AlbyHojel

a month ago

Scale vs Architecture Question: if you want to build a simulation engine using a video gen model, should you incorporate inductive biases that promote 3D consistency? Bill from the Sora team’s answer: inductive biases always come back to bite you. Scale is enough.

sedielem Profile Picture

Sander Dieleman @sedielem

2 months ago

The way overfitting is usually taught: you underfit for a while, then at some point, you start overfitting. This "phase transition" perspective can be misleading. As Alex points out, you can have both at the same time. It's probably more useful to think of it as a trade-off.

unixpickle Profile Picture

Alex Nichol @unixpickle

2 months ago

The way overfitting is usually taught: you underfit for a while, then at some point, you start overfitting. This "phase transition" perspective can be misleading. As Alex points out, you can have both at the same time. It's probably more useful to think of it as a trade-off. https://t.co/XM49OGrSQP

Sander Dieleman retweeted
emiel_hoogeboom Profile Picture

Emiel Hoogeboom @emiel_hoogeboom

2 months ago

Want to sample fast from diffusion models? Check out our work on multistep consistency. It turns out that training consistency models over multiple sections is much easier than over one big one. Even I can do it. For more detail see thread below (1/7)

JonathanHeek Profile Picture

Jonathan Heek @JonathanHeek

2 months ago

Want to sample fast from diffusion models? Check out our work on multistep consistency. It turns out that training consistency models over multiple sections is much easier than over one big one. Even I can do it. For more detail see thread below (1/7)

Sander Dieleman retweeted
CSProfKGD Profile Picture

Kosta Derpanis @CSProfKGD

2 months ago

❤️ this frequency interpretation of diffusion models for natural signals as progressively obscuring structure (high-to-low frequency) in the forward noising process and adding structure in the generative reverse process (low-to-frequency). Shoutout to @sedielem and @ArashVahdat…

CSProfKGD Profile Picture

Kosta Derpanis @CSProfKGD

7 months ago

❤️ this frequency interpretation of diffusion models for natural signals as progressively obscuring structure (high-to-low frequency) in the forward noising process and adding structure in the generative reverse process (low-to-frequency). Shoutout to @sedielem and @ArashVahdat… https://t.co/iW83lnTSzj

Sander Dieleman retweeted
WillSmith2real Profile Picture

Will Smith @WillSmith2real

2 months ago

This is getting out of hand! - Will Smith

Sander Dieleman retweeted
jaschasd Profile Picture

Jascha Sohl-Dickstein @jaschasd

3 months ago

Have you ever done a dense grid search over neural network hyperparameters? Like a *really dense* grid search? It looks like this (!!). Blueish colors correspond to hyperparameters for which training converges, redish colors to hyperparameters for which training diverges.

sedielem Profile Picture

Sander Dieleman @sedielem

3 months ago

This one's easy! That honour goes to "the diffusion bible", as I like to call it. It's been well over a year and I still refer to it several times a week. Very few papers I've read come close, in terms of signal-to-noise ratio. arxiv.org/abs/2206.00364

sp_monte_carlo Profile Picture

Sam Power @sp_monte_carlo

3 months ago

This one's easy! That honour goes to "the diffusion bible", as I like to call it. It's been well over a year and I still refer to it several times a week. Very few papers I've read come close, in terms of signal-to-noise ratio. arxiv.org/abs/2206.00364 https://t.co/gwzPodcJV2

Sander Dieleman retweeted
docmilanfar Profile Picture

Peyman Milanfar @docmilanfar

3 months ago

the math in most diffusion papers

Sander Dieleman retweeted
AravSrinivas Profile Picture

Aravind Srinivas @AravSrinivas

3 months ago

A misconception, also sometimes polled by @elonmusk (for humor, I believe), is a false dichotomy between Transformers and Diffusion. Transformer is a neural network architecture. Diffusion is a particular way of modeling the data distribution. These are two separate things. You…

Trends for United States

#bucciovertimechallenge
3.578 posts
Mike Perry
2.330 posts
Logan Thompson
1.679 posts
Lakers in 7
20,2 B posts
Sulphur
7.549 posts
#TexasHockey
2.790 posts
#VegasBorn
2.772 posts
Colin Jost
18 B posts
#BKFCKM4
Oklahoma
56,3 B posts
#KnucklemaniaIV
1.843 posts
Leafs
33,2 B posts
Marner
13 B posts
Hintz
Chivas
32,2 B posts
Duchene
Ardmore
6.251 posts
Feds
55,5 B posts
Seguin
1.146 posts
George Kirby
2.579 posts

You might like

Cofounded and lead @PyTorch at Meta. Also dabble in robotics at NYU. AI is delicious when it is accessible and open-source.
Soumith Chintala
@soumithchintala
186K Followers
ML with Social Purpose. @shakir@mas.to | Research Scientist @DeepMind | Strengthening African ML @DeepIndaba. He/Him. South African 🇿🇦🏳️‍🌈🌍
Shakir Mohamed
@shakir_za
44K Followers
Google DeepMind researcher, machine learning professor, ex-Twitter Cortex, father of 4, wine/music/comedy enthusiast
Hugo Larochelle
@hugo_larochelle
113K Followers
Reverse engineering neural networks at @AnthropicAI. DMs open! Previously @distillpub, OpenAI Clarity Team, Google Brain. Personal account.
Chris Olah
@ch402
91K Followers
Researcher in Deep Learning @GoogleDeepMind. Angel investor. Co-creator @GoogleAI Brain Amsterdam, Ex @DeepMind, Edu at Oxford, UvA and Stanford.
Nal
@nalkalc
35K Followers
UPMC Professor of Computer Science at Carnegie Mellon University, ex-Director of AI research at @Apple, co-founder Perceptual Machines (acquired by Apple)
Russ Salakhutdinov
@rsalakhu
100K Followers
Staff research scientist at DeepMind. AI and neuro. Former physicist, current human. https://t.co/jVR7w1pUOQ
Jane Wang
@janexwang
36K Followers
VP of Research & Deep Learning Lead, Google DeepMind. Gemini co-lead. Past: AlphaStar, AlphaFold, AlphaCode, WaveNet, seq2seq, distillation, TF.
Oriol Vinyals
@OriolVinyalsML
166K Followers
Machine Learning Researcher, CS Professor (@PrincetonCS), Dad, Woodworker
Ryan Adams
@ryan_p_adams
34K Followers
Director, @StanfordAILab. Assoc. Director, @StanfordHAI. Founder, @stanfordnlp. Prof. CS & Linguistics, @Stanford. IP @aixventureshq. 🇦🇺 Do #NLProc & #AI. 👋
Christopher Manning
@chrmanning
126K Followers
I research intelligence to understand it and to harness it wisely. Path: Wits, Cambridge, Berkeley, UBC, Oxford, DarkBlueLabs, Google DeepMind
Nando de Freitas 🏳..
@NandoDF
97K Followers
Open-Endedness Team Lead @GoogleDeepMind, Professor of AI @AI_UCL, PI @UCL_DARK, @ELLISforEurope Scholar. ex @MetaAI (FAIR), @CompSciOxford. Opinions my own.
Tim Rocktäschel
@_rockt
29K Followers
Associate Professor of Machine Learning, University of Oxford @OATML_Oxford Group Leader Director of Research at AISI (formerly UK Taskforce on Frontier AI)
Yarin
@yaringal
38K Followers
Multilingual LLMs @cohere • Prev: @GoogleDeepMind • Newsletter: https://t.co/7JGh2qpG98
Sebastian Ruder
@seb_ruder
80K Followers
FR/US/GB AI/ML Person, Director of Research at @GoogleDeepMind, Honorary Professor at @UCL_DARK, @ELLISforEurope Fellow. All posts are personal.
Edward Grefenstette
@egrefen
36K Followers