Papers

Variational Autoencoders for Efficient Simulation-Based Inference

M Nautiyal, A Shternshis, A Hellander… - arXiv preprint arXiv …, 2024 - arxiv.org
Computer Science paper cs.LG Suggest

… Contribution: We propose a variational inference approach for simulation-based inference, utilizing a Conditional Variational Autoencoder (C-VAE) architecture [Kingma …

Link to paper

BibTeX

@article{2411.14511v1,
Author        = {Mayank Nautiyal and Andrey Shternshis and Andreas Hellander and Prashant Singh},
Title         = {Variational Autoencoders for Efficient Simulation-Based Inference},
Eprint        = {2411.14511v1},
ArchivePrefix = {arXiv},
PrimaryClass  = {cs.LG},
Abstract      = {We present a generative modeling approach based on the variational inference
framework for likelihood-free simulation-based inference. The method leverages
latent variables within variational autoencoders to efficiently estimate
complex posterior distributions arising from stochastic simulations. We explore
two variations of this approach distinguished by their treatment of the prior
distribution. The first model adapts the prior based on observed data using a
multivariate prior network, enhancing generalization across various posterior
queries. In contrast, the second model utilizes a standard Gaussian prior,
offering simplicity while still effectively capturing complex posterior
distributions. We demonstrate the efficacy of these models on well-established
benchmark problems, achieving results comparable to flow-based approaches while
maintaining computational efficiency and scalability.},
Year          = {2024},
Month         = {Nov},
Url           = {http://arxiv.org/abs/2411.14511v1},
File          = {2411.14511v1.pdf}
}

Share