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

Дата публикации: 17-08-2026 20:26:00

Implicit generative models are often trained adversarially, which can yield unstable dynamics and mode collapse. The invariant statistical loss (ISL) offers a fully sample-based alternative by comparing empirical ranks of real and generated samples. In this work, we formally characterize ISL as a proper divergence over continuous distributions and establish key regularity properties, showing that it is continuous and differentiable, thereby enabling stable gradient-based optimization without adversarial games. We further enhance ISL along two practical axes. First, to better model heavy-tailed data, where Gaussian latent priors can limit tail expressivity, we introduce Pareto-ISL, which replaces Gaussian noise with a generalized Pareto latent distribution to improve the representation of both typical and extreme events. Second, to handle multivariate data at scale, we propose ISL-slicing: a computationally efficient procedure that projects samples onto random one-dimensional subspaces, computes rank-based losses per projection, and averages them to capture high-dimensional structure. Experiments demonstrate improved tail fidelity with Pareto-ISL and show that ISL-slicing scales effectively to high dimensions. Specifically, in high dimensional settings we show that ISL can be used either as a standalone criterion or as a strong pretraining objective for subsequent adversarial fine-tuning.

Схожие новости

#	Наименование новости	Тональность	Информативность	Дата публикации
1	Nonparametric Estimation of a Factorizable Density using Diffusion Models	0	8.59	17-08-2026
2	Generative Bayesian Inference with GANs	0	6.62	17-08-2026
3	Semi-supervised learning for linear extremile regression	0	6.83	17-08-2026
4	Statistical Learning Theory for Neural Operators	0	10.21	17-08-2026
5	Mixing times of data-augmentation Gibbs samplers for high-dimensional probit regression	0	8.78	17-08-2026
6	Knowledge Cascade: Reverse Knowledge Distillation on Nonparametric Multivariate Functional Estimation	0	5.75	17-08-2026
7	High-Dimensional Analysis of Gradient Flow for Extensive-Width Quadratic Neural Networks	0	8.7	17-08-2026
8	Near-optimal Delta-convex Estimation of Lipschitz Functions	0	9.71	17-08-2026
9	Graph-based Clustering Revisited: A Relaxation of Kernel k-Means Perspective	0	10.94	17-08-2026
10	Identifying Weight-Variant Latent Causal Models	0	5.7	17-08-2026

Классификация: Наука. Схожих патентов: 0. Схожих новостей: 10. Тональность: 0. Информативность: 10. Источник: jmlr.org.