This week in AI just changed everything. Massive announcements from OpenAI, Warren Buffet, Microsoft, AMD, Palantir, Tesla, Meta, Humane, IBM, Wendy's, Google, HuggingFace, Scale AI, Synthesia, Anthropic, and Stability AI. EVERYTHING you need to know and why it's important:

The fractional Laplacian is a differential operator generalizing the Laplacian. They are the only rotation/translation/scale invariant linear operator. en.wikipedia.org/wiki/Fractiona…

Optimization can be carried over on manifolds using gradient descent, where the manifold gradient is the projection of an extrinsic gradient on the tangent plane. sites.uclouvain.be/absil/amsbook/

Looking forward to AIM 2023 (AI & Mathematics) this week June 1-2 at our University University of Twente (inactive). Final chance to register today to enjoy great talks and connect. --> aimath.nl/index.php/even… #ai #mathematics #aim #NWO #utwente #dsi

Fourier is to convolution what Legendre is to inf-convolution. en.wikipedia.org/wiki/Fourier_t… en.wikipedia.org/wiki/Legendre_…

We are releasing code for our algorithm that won the 2021 MICCAI/SMRA carotid artery segmentation challenge in black blood MRI: github.com/MIAGroupUT/car… 📜Paper by Dieuwertje Alblas: doi.org/10.1117/12.261… Many thanks to @ElinaSutre for setting up the repo and docs! Christoph Brune

Happy to share that our paper "What do neural networks learn in image classification? A frequency shortcut perspective" is accepted at #ICCV2023 Proud of my #phd student Shunxin Wang, who led the project (collab w/ Christoph Brune and Raymond Veldhuis, University of Twente (inactive) ), cu in Paris!

Handing the flag onto Holland to organise the next ICIAM. Tokyo has set a very high standard to follow! Sad it's all over. It was a great meeting!!!

Still present at #MICCAI2023 in Vancouver? Come drop by our poster during lunch, where I will gladly tell you all about automatic quality assessment for segmentation in large datasets! Joint work with Elina Thibeau-Sutre, Sophie Buurman, Jelmer Wolterink Christoph Brune

The best inventions of 2023? TIME just released their list highlighting the year's top innovations. The 14 AI tools that made the list:

We are looking for excellent candidates for a PhD position at University of Twente (inactive) with Jelmer Wolterink and me in close collaboration with Siemens Healthineers. The project focuses on model-driven DL for prostate MRI analysis. More info and application before Nov 17 here: utwentecareers.nl/en/vacancies/1…

What do polar coordinates, polar matrix factorization, & Helmholz decomposition of a vector field have in common? They’re all implied by Brenier’s Theorem: a cornerstone of Optimal Transport theory. It’s a fundamental decomposition result & deserves to be better known. 1/5

After 3 years, it's time for us to start sharing the chapters of the GDL book! ❤️ Also included: companion slides from our Cambridge University & University of Oxford courses 🧑‍🎓 Chapter 1 is out **now**! More to follow soon 🎉 geometricdeeplearning.com/book 📖 Michael Bronstein @joanbruna Taco Cohen

Great plenary talk by Sjoerd Dirksen: telling us about the separation and memorization capacity of randomly initialized neural networks at the 3rd AI & Math workshop at Utrecht University. Christoph Brune

🚀 Day 2 Oral Session 4 Enhancing Low-Quality Medical Images presentation at MIUA 2024 #MIUA2024 #Conference #researchpaper #Innovation

Truly excellent piece on entropy. Source: Quanta Magazine search.app/ichMj5mCxENsWC…

Enschede crowd you don't want to miss this! Join us for two exciting talks by Jolanda Wentzel and Erik Bekkers. As a bonus, you get to see me in "appropriate traditional dress from [my] country of origin". 📅 28 May 📍 UTwente campus

Paper: • sciencedirect.com/science/articl… • arxiv.org/abs/2410.11920 Code: • github.com/sukjulian/deep… Collaboration with Guido Nannini, Patryk Rygiel, Christoph Brune, Gianluca Pontone, Alberto Redaelli and Jelmer Wolterink

New art project. Train and inference GPT in 243 lines of pure, dependency-free Python. This is the *full* algorithmic content of what is needed. Everything else is just for efficiency. I cannot simplify this any further. gist.github.com/karpathy/8627f…

this is really amazing. Andrej implemented most of the complexity of modern LLMs in <200 lines. and only imported these three packages