HyperCore: Coreset Selection under Noise via Hypersphere Models

Brian Bernhard Moser, Arundhati Shanbhag, Tobias Christian Nauen, Stanislav Frolov, Federico Raue, Joachim Folz, Andreas Dengel

Keywords: Coreset Selection, Label Noise, Hypersphere Models, Adaptive Data Pruning

TL;DR: HyperCore is a robust coreset selection method that uses lightweight, per-class hypersphere models to distinguish clean from noisy data, automatically learning an optimal pruning threshold to create compact datasets without requiring a fixed budget.

Abstract:

The goal of coreset selection methods is to identify representative subsets of datasets for efficient model training.
Yet, existing methods often ignore the possibility of annotation errors and require fixed pruning ratios, making them impractical in real-world settings.
We present HyperCore, a robust and adaptive coreset selection framework designed explicitly for noisy environments.
HyperCore leverages lightweight hypersphere models learned per class, embedding in-class samples close to a hypersphere center while naturally segregating out-of-class samples based on their distance.
By using Youden’s J statistic, HyperCore can adaptively select pruning thresholds, enabling automatic, noise-aware data pruning without hyperparameter tuning.
Our experiments reveal that HyperCore consistently surpasses state-of-the-art coreset selection methods, especially under noisy and low-data regimes.
HyperCore effectively discards mislabeled and ambiguous points, yielding compact yet highly informative subsets suitable for scalable and noise-free learning.
The code for HyperCore will be published upon acceptance.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Submission Number: 2361