Does a smaller latent space lead to worse generation in latent diffusion models? Not necessarily! We show that LDMs are extremely robust to a wide range of compression rates (10-1000x) in the context of physics emulation.

We got lost in latent space. Join us 👇

Thank you Jonah, you did a great job as well 😊

The Well was featured in a LANL press release! This is a huge dataset designed for training scientific #AI and foundation models. The #HPC simulations of neutron star mergers, some of the most energetic events in the universe, we're my contribution. ⚛️🔭🧪👇
www.lanl.gov/media/news/1...

I will be attending #NeurIPS2024! Feel free to ping me or the @polymathicai.bsky.social team if you want to chat about The Well 🥤or MPP 🚤! And since you like eye-catchy scientific gifs, here is a Rayleigh-Bénard convection one 😁
🥤: arxiv.org/abs/2412.00568
🚤: arxiv.org/abs/2310.02994

Amazing work! A big physics-informed ML dataset with actually relevant problems, created in cooperation with domain experts. It‘s time to finally move on from 1D Burgers‘ equations 🚀

🧵 Today with @polymathicai.bsky.social and others we're releasing two massive datasets that span dozens of fields - from bacterial growth to supernova!

We want this to enable multi-disciplinary foundation model research.

Generating cat videos is nice, but what if you could tackle real scientific problems with the same methods? 🧪🌌
Introducing The Well: 16 datasets (15TB) for Machine Learning, from astrophysics to fluid dynamics and biology.
🐙: github.com/PolymathicAI...
📜: openreview.net/pdf?id=00Sx5...

Today along with my project co-lead @rubenohana.bsky.social and the team at @polymathicai.bsky.social I'm excited to announce the release of the Well, a 15TB collection of 15+ datasets for physical simulation.

Paper: openreview.net/pdf?id=00Sx5...
Github: github.com/PolymathicAI...

Deep learning can do more than generating cute cat pictures. It is a game changer for Science as a whole. But benchmarking DL methods on scientific data is complex ... until today!

I am super proud to have contributed to The Well as part of my internship at @polymathicai.bsky.social !

Links:
📜 #NeurIPS2024: openreview.net/pdf?id=00Sx5...
🐙 GitHub: github.com/PolymathicAI...
🤗 Blog: huggingface.co/blog/rubenoh...
💻 PyPI: pypi.org/project/the-...

Shout out to the @polymathicai.bsky.social team for the great work. What's next is exciting, stay tuned!

🌌 Bridging the gap between ML & numerical simulations!

With The Well, ML researchers can tackle advanced physics, and simulation experts explore efficient surrogates. Imagine predicting the explosion of supernova in seconds! 🌠
Let’s accelerate scientific discovery together!🤝

For ML researchers, The Well offers:
🧠 Unique benchmarks (spatial-temporal patterns ≠ natural videos)
📐 Many knowledge transfer tasks
⏳ Handling temporal variation
⚙️ Generalizing across physical parameters
This is an opportunity to advance scientific research with ML! 🚀

🤔 Why The Well?
Unlike natural datasets, scientific data is hard to gather & interpret. Evaluating a turbulent astrophysical process isn’t like judging a cat photo! 🐱
We standardized complex scientific datasets to let ML researchers focus on what matters: predicting physics. 📈

Generating cat videos is nice, but what if you could tackle real scientific problems with the same methods? 🧪🌌
Introducing The Well: 16 datasets (15TB) for Machine Learning, from astrophysics to fluid dynamics and biology.
🐙: github.com/PolymathicAI...
📜: openreview.net/pdf?id=00Sx5...

Would love to be added to, thanks!

Researchers in NYC👇

Hello, researcher in NYC on foundation models for science 👋

Research fellowships at the Flatiron Institute's Center for Computational Mathematics in NYC: apply.interfolio.com/155357

Machine learning @ CCM is growing! users.flatironinstitute.org/~lsaul/ml_cc...

We work with many others in the sciences across FI: www.simonsfoundation.org/flatiron/mac...

Hello World!