Multi-Label Gold Asymmetric Loss Correction with Single-Label Regulators

Multi-Label Gold Asymmetric Loss Correction with Single-Label Regulators

Pene, Cosmin (TU Delft Electrical Engineering, Mathematics and Computer Science; TU Delft Software Technology)

Chen, Lydia Y. (mentor)
Ghiassi, S. (mentor)
Younesian, T. (mentor)
Kuipers, F.A. (graduation committee)

Delft University of Technology

Computer Science and Engineering

CSE3000 Research Project

2021-07-02

Multi-label learning is an emerging extension of the multi-class classification where an image contains multiple labels. Not only acquiring a clean and fully labeled dataset in multi-label learning is extremely expensive, but also many of the actual labels are corrupted or missing due to the automated or non-expert annotation techniques. Noisy label data decrease the prediction performance drastically. In this paper, we propose a novel Gold Asymmetric Loss Correction with Single-Label Regulators (GALC-SLR) that operates robust against noisy labels. GALC-SLR estimates the noise confusion matrix using single-label samples, then constructs an asymmetric loss correction via estimated confusion matrix to avoid overfitting to the noisy labels. Empirical results show that our method outperforms the state-of-the-art original asymmetric loss multi-label classifier under all corruption levels, showing mean average precision improvement up to 28.67\% on a real-world dataset of MS-COCO, yielding a better generalization of the unseen data and increased prediction performance.

Multi-label classification
Label Noise
Robust learning
Deep Neural Networks
Noise estimation
Asymmetric Loss

http://resolver.tudelft.nl/uuid:b42c6372-ea80-4b1a-b3ee-fa00f4ad8998

Student theses

bachelor thesis

© 2021 Cosmin Pene