top of page

The Full Story

Research Interests 

Our group pioneers in the application and development of new deep learning methods for astrophysical challenges. We have a particular love for fundamental questions about our Universe, but also have dived into other problems ranging from genomics to climate research.

Interpretable Learning for Science

We figured out how to combine neural networks and genetic programming to understand what neural networks have learnt from the data. 

In gist, we figured out how to extract what the AI has learnt from observing our world. 

Animation credit: Miles Cranmer

Simulation Based Inference

We try to understand the hidden parameters  of our own Universe by comparing our observed Universe to the simulated Universe. 

Stay Tuned

Making the biggest map of our Universe

Our group analyzes the observations of the sky using telescopes both on the ground and in space in a wide range of wavelengths. We analyze these data to try and understand the beginning, the content and possibly the end of the Universe

Simulations Acceleration by AI 

In order to understand our Universe, we need very very fast simulation of our Universe as our "theoretical prediction". Very often, these simulations are too slow to be used for inferring what is in our Universe. So we change this with AI.

Animation credit: Illustris Project

bottom of page