Maximum Likelihood Learning of Energy-Based Models for Simulation-Based Inference - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail (Preprint/Prepublication) Année : 2022

Maximum Likelihood Learning of Energy-Based Models for Simulation-Based Inference

Pierre Glaser
  • Fonction : Auteur
  • PersonId : 1118675
Arnaud Doucet
  • Fonction : Auteur
  • PersonId : 1192870
Arthur Gretton
  • Fonction : Auteur
  • PersonId : 943040

Résumé

We introduce two synthetic likelihood methods for Simulation-Based Inference (SBI), to conduct either amortized or targeted inference from experimental observations when a high-fidelity simulator is available. Both methods learn a conditional energy-based model (EBM) of the likelihood using synthetic data generated by the simulator, conditioned on parameters drawn from a proposal distribution. The learned likelihood can then be combined with any prior to obtain a posterior estimate, from which samples can be drawn using MCMC. Our methods uniquely combine a flexible Energy-Based Model and the minimization of a KL loss: this is in contrast to other synthetic likelihood methods, which either rely on normalizing flows, or minimize score-based objectives; choices that come with known pitfalls. Our first method, Amortized Unnormalized Neural Likelihood Estimation (AUNLE), introduces a tilting trick during training that allows to significantly lower the computational cost of inference by enabling the use of efficient MCMC techniques. Our second method, Sequential UNLE (SUNLE), employs a robust doubly intractable approach in order to re-use simulation data and improve posterior accuracy on a specific dataset. We demonstrate the properties of both methods on a range of synthetic datasets, and apply them to a neuroscience model of the pyloric network in the crab Cancer Borealis, matching the performance of other synthetic likelihood methods at a fraction of the simulation budget.
Fichier principal
Vignette du fichier
2210.14756.pdf (7.27 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03869080 , version 1 (24-11-2022)

Identifiants

Citer

Pierre Glaser, Michael Arbel, Arnaud Doucet, Arthur Gretton. Maximum Likelihood Learning of Energy-Based Models for Simulation-Based Inference. 2022. ⟨hal-03869080⟩
42 Consultations
24 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More