0

Training Energy-Based Normalizing Flow with Score-Matching Objectives

A new flow-based generative model, EBFlow, uses score-matching objectives to avoid computing Jacobian determinants for linear transformations, improving training efficiency and empirical performance.

Year
2023
Venue
training-energy-based-normalizing-flow-with
Authors
5
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2305.15267v2ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

Abstract

In this paper, we establish a connection between the parameterization of flow-based and energy-based generative models, and present a new flow-based modeling approach called energy-based normalizing flow (EBFlow). We demonstrate that by optimizing EBFlow with score-matching objectives, the computation of Jacobian determinants for linear transformations can be entirely bypassed. This feature enables the use of arbitrary linear layers in the construction of flow-based models without increasing the computational time complexity of each training iteration from $O(D^2L)$ to $O(D^3L)$ for an $L$-layered model that accepts $D$-dimensional inputs. This makes the training of EBFlow more efficient than the commonly-adopted maximum likelihood training method. In addition to the reduction in runtime, we enhance the training stability and empirical performance of EBFlow through a number of techniques developed based on our analysis of the score-matching methods. The experimental results demonstrate that our approach achieves a significant speedup compared to maximum likelihood estimation while outperforming prior methods with a noticeable margin in terms of negative log-likelihood (NLL).

Authors

5