LINFA: a Python library for variational inference with normalizing flow and annealing

TL;DR

LINFA is a Python library that enhances variational inference by integrating normalizing flows and annealing techniques, enabling efficient approximation of complex, computationally expensive models with dependent parameters.

Contribution

The paper introduces LINFA, a novel Python library that combines normalizing flows and annealing for improved variational inference in challenging models.

Findings

LINFA performs well on various benchmarks.

It effectively handles complex, high-dimensional distributions.

The library is publicly available for use and further development.

Abstract

Variational inference is an increasingly popular method in statistics and machine learning for approximating probability distributions. We developed LINFA (Library for Inference with Normalizing Flow and Annealing), a Python library for variational inference to accommodate computationally expensive models and difficult-to-sample distributions with dependent parameters. We discuss the theoretical background, capabilities, and performance of LINFA in various benchmarks. LINFA is publicly available on GitHub at https://github.com/desResLab/LINFA.