Paper: www.nature.com/articles/s41...

GitHub: github.com/artificial-s...

Press release: mpl.mpg.de/news-events/...

Our work constitutes an important step in AI-driven scientific discovery of new concepts in optics and advanced microscopy.

This couldn't have been possible without the amazing team at the Max Planck Institute for the Science of Light!

w/ Sören Arlt, Leonhard Möckl and @mariokrenn.bsky.social

8/

To validate our approach, we first had XLuminA rediscover groundbreaking techniques like STED microscopy (2014 Nobel Prize in Chemistry).

But it went further — finding a completely new optical configuration that combines multiple SR principles in unexpected ways! 👀

7/

XLuminA is written in JAX and features seamlessly integrated automatic differentiation and GPU-acceleration.

We demonstrate a speed-up of x10,000 compared to traditional optimization methods! 💥🚀

6/

Our solution: we have developed XLuminA, an efficient open-source framework designed with the ultimate goal of discovering new experimental optical designs, with a focus on SR microscopy.

5/

Thus, AI-based exploration techniques could provide enormous benefit, by exploring this space in a fast, unbiased way! 🤖🌌

4/

However, the vast space encompassing all possible experimental configurations in optics suggests that some powerful designs might have not been discovered yet, and might never be with a direct human-driven approach. 🤔

3/

Finding new microscopy techniques has traditionally relied on human experience, intuition and creativity. The discovery of super-resolution methods, which circumvent the classical diffraction limit of light, represent a leap in optical microscopy.

2/

Happy to see our work in its published version at @naturecomms.bsky.social !

We developed an AI framework that autonomously discovers new experimental designs in super-resolution microscopy 🔬🤖✨

1/🧵