Statistics - Machine Learning; Computer Science - Machine Learning; High Energy Physics - Phenomenology; Physics - Data Analysis; Statistics and Probability
Abstract :
[en] We extend recent work (Brehmer, et. al., 2018) that use neural networks as surrogate models for likelihood-free inference. As in the previous work, we exploit the fact that the joint likelihood ratio and joint score, conditioned on both observed and latent variables, can often be extracted from an implicit generative model or simulator to augment the training data for these surrogate models. We show how this augmented training data can be used to provide a new cross-entropy estimator, which provides improved sample efficiency compared to previous loss functions exploiting this augmented training data.
Disciplines :
Physics Computer science
Author, co-author :
Stoye, Markus
Brehmer, Johann
Louppe, Gilles ; Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Big Data
Pavez, Juan
Cranmer, Kyle
Language :
English
Title :
Likelihood-free inference with an improved cross-entropy estimator
Publication date :
02 August 2018
Event name :
Machine Learning and the Physical Sciences, NeurIPS 2019