Papers

Simulation-based Inference with the Generalized Kullback-Leibler Divergence

BK Miller, M Federici, C Weniger, P Forré - arXiv preprint arXiv …, 2023 - arxiv.org
Statistics paper stat.ML Suggest

… In Simulation-based Inference, the goal is to solve the inverse problem when the likelihood is only known implicitly. Neural Posterior Estimation commonly fits a normalized …

Link to paper

BibTeX

@article{2310.01808v1,
Author        = {Benjamin Kurt Miller and Marco Federici and Christoph Weniger and Patrick Forré},
Title         = {Simulation-based Inference with the Generalized Kullback-Leibler
  Divergence},
Eprint        = {2310.01808v1},
ArchivePrefix = {arXiv},
PrimaryClass  = {stat.ML},
Abstract      = {In Simulation-based Inference, the goal is to solve the inverse problem when
the likelihood is only known implicitly. Neural Posterior Estimation commonly
fits a normalized density estimator as a surrogate model for the posterior.
This formulation cannot easily fit unnormalized surrogates because it optimizes
the Kullback-Leibler divergence. We propose to optimize a generalized
Kullback-Leibler divergence that accounts for the normalization constant in
unnormalized distributions. The objective recovers Neural Posterior Estimation
when the model class is normalized and unifies it with Neural Ratio Estimation,
combining both into a single objective. We investigate a hybrid model that
offers the best of both worlds by learning a normalized base distribution and a
learned ratio. We also present benchmark results.},
Year          = {2023},
Month         = {Oct},
Url           = {http://arxiv.org/abs/2310.01808v1},
File          = {2310.01808v1.pdf}
}

Share