Set-Level Self-Supervised Learning from Noisily-Labeled Data
Lists
Only you see this
Abstract
Noisy labels are inevitably presented in real-world datasets due to labeling error or visual content ambiguity. Existing methods generally approach the task of noisy label learning (NLL) by either properly regularizing the model, or reweighting clean/noisy labeled samples. While self-supervised learning (SSL) has been applied to pre-train deep neural networks without label supervision, downstream tasks like image classification still require clean labeled data. And, most SSL strategies are performed at the instance level, regardless of the correctness of its label. In this paper, we propose set-level self-supervised learning (SLSSL), which performs SSL at mini-batch levels with observed noisy labels. By corrupting the labels of each training mini-batch, our SLSSL enforces the model to exhibit sufficient robustness. Moreover, the proposed SLSSL can also be utilized for sample reweighting technique. As a result, the proposed learning scheme can be applied as an expectation-maximization (EM) algorithm during model training. Extensive experiments on synthetic and real-world noisy label data confirm the effectiveness of our framework.
BibTeX
@inproceedings{lin2023set,
author = {Chia-Ching Lin and Shang-Fu Chen and Fu-En Yang and Yu-Chiang Frank Wang and Chin-Laung Lei},
title = {{Set-Level Self-Supervised Learning from Noisily-Labeled Data}},
booktitle = {International Conference on Learning Representations},
year = {2023},
url = {https://openreview.net/forum?id=xeg2fW5E2l3},
}Loading related content...