Machine Learning Seminar Series | Accelerating fundamental physics: From theory design to experimental tests at CERN

Abstract: Fundamental physics is about designing bottom-up theories to explain the Universe, from elementary particles to the nature of gravity. Theoretical progress has traditionally relied on human insight into complex mathematics, with no systematic way to explore the space of ideas. I will discuss a series of neuro-symbolic AI results that systematize this approach while minimizing biases to prior literature. Our AI tools also compute theory predictions and help design experimental tests. These AI tools also help calculate theory predictions and design experimental tests while providing human-readable insights from problem-solving process and explicitly cite previous scientific literature that inspired new LLM-generated ideas.

On the experimental side, testing theories at the world's largest experiments, such as the Large Hadron Collider, is yet another computational challenge. High-dimensional data arrive at 40 MHz (~40 TB/s), and decisions must be made rapidly. I will discuss recent advancements in robust neural inference techniques with reliable uncertainty quantification that let us exploit the full high-dimensional data in hypothesis tests. Additional machine learning methods make these genAI predictions trustworthy, enabling tests that outperform likelihood-ratio-based statistical tests.

Bio: Aishik Ghosh is an assistant professor in the School of Physics at Georgia Tech with a focus on developing AI methods to accelerate fundamental physics and astrophysics. His group works on theoretical physics, statistical methods, and experiment design. For robust scientific applications, Dr. Ghosh focuses on uncertainty quantification, interpretability, and verifiability of AI algorithms, targeting publications in physics journals and ML conferences. He has also published policy documents with the Organization for Economic Co-operation and Development and took a leading role in a book on AI and the Future of Science. He received his PhD in particle physics from Paris-Saclay University for introducing digital twins and neural inference techniques to a Large Hadron Collider experiment and was a postdoc at UC Irvine and Lawrence Berkeley National Laboratory. Website: https://profaishikghosh.github.io/

For CODA guest access, please contact shatcher8@gatech.edu at least 2 business days prior to the event.

Join Via Zoom: https://gatech.zoom.us/j/97583133407?pwd=arlGKeB48JGRzqOQt3mPELCH1sfL5J.1 

Meeting ID: 975 8313 3407 

Passcode: 799040