Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Density Deconvolution with Normalizing Flows
(Submitted on 16 Jun 2020 (v1), last revised 13 Jul 2020 (this version, v2))
Abstract: Density deconvolution is the task of estimating a probability density function given only noise-corrupted samples. We can fit a Gaussian mixture model to the underlying density by maximum likelihood if the noise is normally distributed, but would like to exploit the superior density estimation performance of normalizing flows and allow for arbitrary noise distributions. Since both adjustments lead to an intractable likelihood, we resort to amortized variational inference. We demonstrate some problems involved in this approach, however, experiments on real data demonstrate that flows can already out-perform Gaussian mixtures for density deconvolution.
Submission history
From: James Ritchie [view email][v1] Tue, 16 Jun 2020 18:00:04 GMT (1045kb,D)
[v2] Mon, 13 Jul 2020 10:58:53 GMT (1045kb,D)
Link back to: arXiv, form interface, contact.