What type of study is this?

This is a Experimental Study study.

October 11, 2025Open Access

Disentangled Deep Smoothed Bootstrap for Fair Imbalanced Regression

SSSamuel StocksiekerFondation de France DPDenys PommeretCentre National de la Recherche Scientifique ACArthur CharpentierCentre National de la Recherche Scientifique

Key Points

The proposed method significantly improves predictive performance on imbalanced regression tasks, addressing a common challenge in modeling.
Using a novel combination of disentangled variational autoencoders and smoothed bootstrap, the approach generates improved latent representations of imbalanced data.
Numerical comparisons demonstrate the method's superior efficiency over traditional approaches on benchmark datasets focused on imbalanced regression.
The findings highlight the importance of tailored methods for imbalanced regression problems, suggesting broader applications in predictive modeling.

Abstract

Imbalanced distribution learning is a common and significant challenge in predictive modeling, often reducing the performance of standard algorithms. Although various approaches address this issue, most are tailored to classification problems, with a limited focus on regression. This paper introduces a novel method to improve learning on tabular data within the Imbalanced Regression (IR) framework, which is a critical problem. We propose using Variational Autoencoders (VAEs) to model and define a latent representation of data distributions. However, VAEs can be inefficient with imbalanced data like other standard approaches. To address this, we develop an innovative data generation method that combines a disentangled VAE with a Smoothed Bootstrap applied in the latent space. We evaluate the efficiency of this method through numerical comparisons with competitors on benchmark datasets for IR.

KI fragen

Bookmark

View Full Paper