Approximating Likelihood Ratios with Calibrated Discriminative Classifiers

Preprint English OPEN
Cranmer, Kyle; Pavez, Juan; Louppe, Gilles;
  • Subject: : Physics [Physical, chemical, mathematical & earth Sciences] | : Physique [Physique, chimie, mathématiques & sciences de la terre] | Statistics - Machine Learning | 62P35 | Statistics - Applications | Physics - Data Analysis, Statistics and Probability | : Computer science [Engineering, computing & technology] | 62P35, 62F99, 62H30 | : Sciences informatiques [Ingénierie, informatique & technologie] | 62H30 | Statistics and Probability | 62F99 | Physics - Data Analysis

In many fields of science, generalized likelihood ratio tests are established tools for statistical inference. At the same time, it has become increasingly common that a simulator (or generative model) is used to describe complex processes that tie parameters $\theta$ o... View more