Data Science Seminar Series: Gilles Louppe

on the November 19, 2018

17:00 to 18:00
The Data Science Seminar Series bring high profile researchers to Université Grenoble Alpes to show the increasing role of data science in modern research. It is open to all researchers of Université Grenoble Alpes. For the 3rd seminar of this year, we will welcome Gilles Louppe from University of Liège.
Gilles Louppe from University of Liège will give a talk on “Likelihood-free inference for Physical Sciences”. The event will take place on Monday 19 November at IM2AG building. The event will be followed by a coffee break.


Abstract

Simulators often provide the best description of real-world phenomena. However, they also lead to challenging inverse problems because the density they implicitly define is often intractable. We present a new suite of simulation-based inference techniques that go beyond the traditional Approximate Bayesian Computation approach, which struggles in a high-dimensional setting, and extend methods that use surrogate models based on neural networks. We show that additional information, such as the joint likelihood ratio and the joint score, can often be extracted from simulators and used to augment the training data for these surrogate models. Finally, we demonstrate that these new techniques are more sample efficient and provide higher-fidelity inference than traditional methods. 

Related document
Likelihood-free inference for Physical Sciences

Related articles
Adversarial Variational Optimization of Non-Differentiable Simulators
Mining gold from implicit models to improve likelihood-free inference

Published on November 13, 2018

Practical informations

Lieu(x)


Bâtiment IM2AG, bulding F, Amphi 18, campus universitaire de Grenoble