keyword

neural autoregressive flows

Neural autoregressive flows are a class of deep generative models in machine learning that transform simple probability distributions into complex data distributions through a sequence of invertible mappings. They generalize standard autoregressive normalizing flows by replacing simple affine transformations with flexible, strictly monotonic univariate neural networks whose parameters are conditioned on previous variables in an autoregressive sequence. Because each univariate transformation is strictly monotonic, the overall multidimensional transformation is guaranteed to be invertible and yields a triangular Jacobian matrix with a computationally tractable determinant. This structure allows neural autoregressive flows to serve as universal approximators for continuous probability distributions, making them highly expressive for modeling complex multimodal distributions while preserving exact likelihood evaluation.

1 item

Normalizing Flows: An Introduction and Review of Current Methods

Ivan Kobyzev, Simon J.D. Prince, Marcus A. Brubaker

OrganizationsBorealis AIYork University

Why you should read this

Explains the mathematical foundations and architectural design principles of normalizing flows, offering a clear guide to how these models achieve exact probability density estimation and efficient sampling.

Normalizing Flows are generative models which produce tractable distributions where both sampling and density evaluation can be efficient and exact. The goal of this survey article is to give a coherent and comprehensive review of the literature around the construction and use of Normalizing Flows for distribution learning. We aim to provide context and explanation of the models, review current state-of-the-art literature, and identify open questions and promising future directions.

Source

https://arxiv.org/pdf/1908.09257

Added

2026-09-24