paper-with-me

Papers

Combating Label Noise With A General Surrogate Model For Sample Selection

2023-10-16 · Chao Liang, Linchao Zhu, Humphrey Shi, Yi Yang

Modern deep learning systems are data-hungry. Learning with web data is one of the feasible solutions, but will introduce label noise inevitably, which can hinder the performance of deep neural networks. Sample selection is an effective way to deal with label noise. The key is to separate clean samples based on some criterion. Previous methods pay more attention to the small loss criterion where small-loss samples are regarded as clean ones. Nevertheless, such a strategy relies on the learning dynamics of each data instance. Some noisy samples are still memorized due to frequently occurring corrupted learning patterns. To tackle this problem, a training-free surrogate model is preferred, freeing from the effect of memorization. In this work, we propose to leverage the vision-language surrogate model CLIP to filter noisy samples automatically. CLIP brings external knowledge to facilitate the selection of clean samples with its ability of text-image alignment. Furthermore, a margin adaptive loss is designed to regularize the selection bias introduced by CLIP, providing robustness to label noise. We validate the effectiveness of our proposed method on both real-world and synthetic noisy datasets. Our method achieves significant improvement without CLIP involved during the inference stage.

📄 PDF Abstract BibTeX arXiv:2310.10463

Code (0)

등록된 구현이 없습니다.

Tasks

MemorizationSelection bias

Methods 이 논문이 사용한 방법론

Adaptive Loss 설명 없음
CLIP Contrastive Language-Image Pre-training (CLIP), consisting of a simplified version of ConVIRT trained from scratch, is an efficient method of image representation learning…

Similar Papers 제목 키워드 기반

Regroup Median Loss for Combating Label Noise

2023-12-11 · Fengpeng Li, Kemou Li, Jinyu Tian, Jiantao Zhou

The deep model training procedure requires large-scale datasets of annotated data. Due to the difficulty of annotating a large number of samples, label noise caused by incorrect annotations is inevitable, resulting in lo…

UNICON: Combating Label Noise Through Uniform Selection and Contrastive Learning

2022-03-28 · CVPR 2022 1 · Nazmul Karim, Mamshad Nayeem Rizve, Nazanin Rahnavard, Ajmal Mian 외

Supervised deep learning methods require a large repository of annotated data; hence, label noise is inevitable. Training with such noisy data negatively impacts the generalization performance of deep neural networks. To…

Contrastive LearningLearning with noisy labelsMemorization

URL: Combating Label Noise for Lung Nodule Malignancy Grading

2023-08-17 · Xianze Ai, Zehui Liao, Yong Xia

Due to the complexity of annotation and inter-annotator variability, most lung nodule malignancy grading datasets contain label noise, which inevitably degrades the performance and generalizability of models. Although re…

Contrastive LearningRelation

Jo-SRC: A Contrastive Approach for Combating Noisy Labels

2021-03-24 · CVPR 2021 1 · Yazhou Yao, Zeren Sun, Chuanyi Zhang, Fumin Shen 외

Due to the memorization effect in Deep Neural Networks (DNNs), training with noisy labels usually results in inferior model performance. Existing state-of-the-art methods primarily adopt a sample selection strategy, whic…

Contrastive LearningMemorization

PNP: Robust Learning From Noisy Labels by Probabilistic Noise Prediction

2022-01-01 · CVPR 2022 1 · Zeren Sun, Fumin Shen, Dan Huang, Qiong Wang 외

Label noise has been a practical challenge in deep learning due to the strong capability of deep neural networks in fitting all training data. Prior literature primarily resorts to sample selection methods for combat…