iiinova — where your digital self keeps evolving

Yasser Souri @yassersouri.bsky.social · 15/07/2025

As expected. Congrats to the authors.

Yasser Souri @yassersouri.bsky.social · 13/07/2025

I remember when I saw the ICML 2015 test of time award winner, I noticed the paper "Learning to Rank Using Gradient Descent" for the first time. Then I got the idea for the "Deep Relative Attributes" paper.

010

Yasser Souri @yassersouri.bsky.social · 13/07/2025

#ICML2025 Test of time award is likely going to the Batch Normalization paper.

000

Yasser Souri @yassersouri.bsky.social · 26/12/2024

A while back when Sam Altman was in India he was asked whether a team with around $10 M could build something to compete with OpenAI and Sam Altman said it was "hopeless". Deepseek-V3 with around $6 M cost for the pre-training run just released a model with very high capability (on benchmarks)

000

Yasser Souri @yassersouri.bsky.social · 16/12/2024

Jason Weston comments on Ilya’s ToT award talk.

000

Yasser Souri @yassersouri.bsky.social · 13/12/2024

Thomas Kipf with the Google IO's DJ bathrobe :D #NeurIPS2024

000

Yasser Souri @yassersouri.bsky.social · 13/12/2024

It is sad to see authors not being able to present their work at #NeurIPS2024 because of visa issues. But some authors went above and beyond. Here is @hadivafaii.bsky.social tele-presenting his work with an impressive setup (ipad, mic, speaker, holder, battery). Well done sir!

0112

Yasser Souri @yassersouri.bsky.social · 12/12/2024

#NeurIPS2024 (@neuripsconf.bsky.social) Day 2 (Wednesday) Experience One of the great things about conferences like NeurIPS is that you get to see people who you admire for different reasons. I also got to see and talk to some. Really happy I got to talk to William Agnew.

100

Yasser Souri @yassersouri.bsky.social · 12/12/2024

Lot's of grokking papers recently. Lol #NeurIPS2024

010

Yasser Souri @yassersouri.bsky.social · 11/12/2024

Sepp Hochreiter claims “the bitter lesson is over”! #neurips2024

0141

Yasser Souri @yassersouri.bsky.social · 11/12/2024

#NeurIPS2024 (@neuripsconf.bsky.social) Day 1 Experience There were a bunch of interesting Tutorial, Talks and events today at NeurIPS. But definitely the highlight of the day was catching up with friends and current and past colleagues and seeing folks.

100

Yasser Souri @yassersouri.bsky.social · 05/12/2024

Ilya Sutskever has won 3 test of time awards at NeurIPS now! 2022: for AlexNet paper 2023: for word2vec paper 2024: for Seq2Seq paper

010

Reposted by Yasser Souri

Arthur Douillard @douillard.bsky.social · 29/11/2024

Excellent explanation of RoPE embedding, from scratch with all the math needed: fleetwood.dev/posts/you-could-have-… And with beautiful 3blue1brown's style of animation: github.com/3b1b/manim. Original RoPE paper: arxiv.org/abs/2104.09864

05310

Reposted by Yasser Souri

NeurIPS Conference @neuripsconf.bsky.social · 22/11/2024

NeurIPS Conference is now Live on Bluesky! -NeurIPS2024 Communication Chairs

1127669

Reposted by Yasser Souri

Daniel Vila @dvilasuero.hf.co · 21/11/2024

Interested in RLHF, DPO, LLM alignment? I've just created this list featuring awesome people like @natolambert.bsky.social . The list is the opposite of exhaustive; I've just joined some days ago 😅 go.bsky.app/MqRGAf2

108519

Yasser Souri @yassersouri.bsky.social · 20/11/2024

For a while I used pipenv and poetry for managing python projects. After a while I tried "uv" and I am never going back. BTW, here is a doc on using uv and pytorch: docs.astral.sh/uv/guides/in...

docs.astral.sh

PyTorch | uv

010

Yasser Souri @yassersouri.bsky.social · 19/11/2024

Borealis AI used to make the coolest t-shirts for ML conferences, they used to have a shirt with “yes we GAN” on it, another one with “ganaca”!

000

Yasser Souri @yassersouri.bsky.social · 19/11/2024

Which paper is going to win the #neurips2024 test of time award? GAN by @ian-goodfellow.bsky.social et. al. or the famous Sequence to Sequence learning paper from Ilya Sutskever and others? Or maybe some other paper which I cannot recall?

110