Addressing Multilabel Imbalance with an Efficiency-Focused Approach Using Diffusion Model-Generated Synthetic Samples

F. Charte, M. A. Dávila, M. D. Pérez-Godoy, M. J. del Jesus · 2025

Read Addressing Multilabel Imbalance with an Efficiency-Focused Approach Using Diffusion Model-Generated Synthetic Samples by F. Charte, M. A. Dávila, M. D. Pérez-Godoy, M. J. del Jesus (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-10822.txt.

Description

Francisco Charte, Miguel Ángel Dávila, María Dolores Pérez-Godoy, and María José del Jesus propose a diffusion model that creates synthetic examples for multilabel learning, in a two thousand twenty-five machine learning paper. Uneven data bias predictive models, and the problem worsens with many output labels.

Resampling with synthetic minority examples is an all-round fix, but quality depends on the generator. Their MultiLabel Diffusion Model adapts image-style diffusion making to multilabel data, filling empty regions of feature space. Tests against several established resampling algorithms show the model holds its own on quality while improving efficiency. The work targets useful classification, ranking, and label distribution tasks facing skewed real-world data sets.

Catalog record

CREATORF. Charte, M. A. Dávila, M. D. Pérez-Godoy, M. J. del Jesus
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE73 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-10822

Browse the full library · About the Tower of Babel