Improving Label Error Detection and Elimination with Uncertainty Quantification

no code implementations • 15 May 2024 • Johannes Jakubik, Michael Vössing, Manil Maskey, Christopher Wölfle, Gerhard Satzger

Therefore, we develop a range of novel, model-agnostic algorithms for Uncertainty Quantification-Based Label Error Detection (UQ-LED), which combine the techniques of confident learning (CL), Monte Carlo Dropout (MCD), model uncertainty measures (e. g., entropy), and ensemble learning to enhance label error detection.

Ensemble Learning Image Classification +2