Robust training of implicit generative models for multivariate and heavy-tailed distributions with an invariant statistical loss

TL;DR

This paper extends the invariant statistical loss (ISL) method to effectively train implicit generative models for multivariate and heavy-tailed distributions, improving stability and tail modeling in high-dimensional data.

Contribution

It introduces Pareto-ISL using GPD noise for heavy tails and a multivariate extension with random projections, enhancing training stability and tail accuracy.

Findings

Pareto-ISL accurately models distribution tails.

Multivariate extension improves high-dimensional data modeling.

Pretraining with ISL reduces mode collapse in GANs.

Abstract

Traditional implicit generative models are capable of learning highly complex data distributions. However, their training involves distinguishing real data from synthetically generated data using adversarial discriminators, which can lead to unstable training dynamics and mode dropping issues. In this work, we build on the \textit{invariant statistical loss} (ISL) method introduced in \cite{de2024training}, and extend it to handle heavy-tailed and multivariate data distributions. The data generated by many real-world phenomena can only be properly characterised using heavy-tailed probability distributions, and traditional implicit methods struggle to effectively capture their asymptotic behavior. To address this problem, we introduce a generator trained with ISL, that uses input noise from a generalised Pareto distribution (GPD). We refer to this generative scheme as Pareto-ISL for conciseness. Our experiments demonstrate that Pareto-ISL accurately models the tails of the distributions while still effectively capturing their central characteristics. The original ISL function was conceived for 1D data sets. When the actual data is $n$-dimensional, a straightforward extension of the method was obtained by targeting the $n$ marginal distributions of the data. This approach is computationally infeasible and ineffective in high-dimensional spaces. To overcome this, we extend the 1D approach using random projections and define a new loss function suited for multivariate data, keeping problems tractable by adjusting the number of projections. We assess its performance in multidimensional generative modeling and explore its potential as a pretraining technique for generative adversarial networks (GANs) to prevent mode collapse, reporting promising results and highlighting its robustness across various hyperparameter settings.