Bayesian Neural Networks for Deep Learning in Cosmology and Gravitational waves

Europe/Paris
APC laboratory

APC laboratory

10 Rue Alice Domon et Léonie Duquet, 75013 Paris
Description

Deep learning attracts more and more interest within the fields of cosmology and gravitational wave astronomy and its potential applications are becoming more and more promising, however so far its usage by physicists has been limited due to its perceived inability to provide statistical uncertainties on the measurements obtained with these methods.

Uncertainties are crucial to evaluate the validity of a scientific result but most physicists using deep learning in their research currently use deterministic networks which are unable to provide an estimation of the error on the result.

A combination of deep learning and bayesian approaches in bayesian neural networks solve this problem by making the networks probabilistic. Over the past few years this field has received great attention from the artificial intelligence community.

This workshop will give the participants the opportunity to learn more about these networks and how to use them for their research. There will be invited speakers from the computer science community who will give lectures and tutorials to teach cosmologists and graviational wave physicists how to take advantage of these tools.

There will be sessions where participants will be invited to share their own experience with deep learning in general and its practical application in their research.

    • 14:00 15:30
      Invited talk

      Introduction to bayesian neural network by invited speaker from computer science community

    • 15:30 16:00
      Coffee break 30m
    • 16:00 18:00
      Contribution talks
    • 09:00 12:30
      Tutorial

      Tutorial following the introduction on bayesian neural networks

    • 14:00 15:30
      Invited talk: Invited talk 2

      Introduction to bayesian neural network by invited speaker from computer science community

    • 15:30 16:00
      Coffee break 30m
    • 16:00 18:00
      Contribution talks
    • 09:00 12:30
      Tutorial: Tutorial 2

      Tutorial following the introduction on bayesian neural networks

Your browser is out of date!

Update your browser to view this website correctly. Update my browser now

×