Integrable Nonparametric Flows

12/03/2020
by   David Pfau, et al.
0

We introduce a method for reconstructing an infinitesimal normalizing flow given only an infinitesimal change to a (possibly unnormalized) probability distribution. This reverses the conventional task of normalizing flows – rather than being given samples from a unknown target distribution and learning a flow that approximates the distribution, we are given a perturbation to an initial distribution and aim to reconstruct a flow that would generate samples from the known perturbed distribution. While this is an underdetermined problem, we find that choosing the flow to be an integrable vector field yields a solution closely related to electrostatics, and a solution can be computed by the method of Green's functions. Unlike conventional normalizing flows, this flow can be represented in an entirely nonparametric manner. We validate this derivation on low-dimensional problems, and discuss potential applications to problems in quantum Monte Carlo and machine learning.

READ FULL TEXT
research
06/15/2020

Why Normalizing Flows Fail to Detect Out-of-Distribution Data

Detecting out-of-distribution (OOD) data is crucial for robust machine l...
research
10/02/2019

Equivariant Flows: sampling configurations for multi-body systems with symmetric energies

Flows are exact-likelihood generative neural networks that transform sam...
research
09/17/2020

1D and 2D Flow Routing on a Terrain

An important problem in terrain analysis is modeling how water flows acr...
research
01/13/2023

Privatitation of Probability Distributions by use Wavelet Integral approach

A naive theory of additive perturbations on a continuous probability dis...
research
07/15/2021

Copula-Based Normalizing Flows

Normalizing flows, which learn a distribution by transforming the data t...
research
02/06/2019

Toward computing sensitivities of average quantities in turbulent flows

Chaotic dynamical systems such as turbulent flows are characterized by a...
research
02/23/2022

Efficient CDF Approximations for Normalizing Flows

Normalizing flows model a complex target distribution in terms of a bije...

Please sign up or login with your details

Forgot password? Click here to reset