• Corpus ID: 219687317

Rethinking the Value of Labels for Improving Class-Imbalanced Learning

@article{Yang2020RethinkingTV,
  title={Rethinking the Value of Labels for Improving Class-Imbalanced Learning},
  author={Yuzhe Yang and Zhi Xu},
  journal={ArXiv},
  year={2020},
  volume={abs/2006.07529}
}
Real-world data often exhibits long-tailed distributions with heavy class imbalance, posing great challenges for deep recognition models. We identify a persisting dilemma on the value of labels in the context of imbalanced learning: on the one hand, supervision from labels typically leads to better results than its unsupervised counterparts; on the other hand, heavily imbalanced data naturally incurs "label bias" in the classifier, where the decision boundary can be drastically altered by the… 
Semi-supervised Long-tailed Recognition using Alternate Sampling
TLDR
This work presents an alternate sampling framework combining the intuitions from successful methods in these two research areas to address the semi-supervised longtailed recognition problem, and demonstrates significant accuracy improvements over other competitive methods on two datasets.
Targeted Supervised Contrastive Learning for Long-Tailed Recognition
TLDR
Targeted supervised contrastive learning (TSC) is proposed, which improves the uniformity of the feature distribution on the hypersphere and achieves state-of-the-art performance on long-tailed recognition tasks.
Towards Resolving the Challenge of Long-tail Distribution in UAV Images for Object Detection
TLDR
The Dual Sampler and Head detection Network (DSHNet) is proposed, which is the first work that aims to resolve long-tail distribution in UAV images and achieves new state-of-the-art performance when combining with image cropping methods.
Relieving Long-tailed Instance Segmentation via Pairwise Class Balance
  • Yin-Yin He, Peizhen Zhang, Xiu-Shen Wei, Xiangyu Zhang, Jian Sun
  • Computer Science
    ArXiv
  • 2022
TLDR
This paper proposes a novel Pairwise Class Balance (PCB) method, built upon a confusion matrix which is updated during training to accumulate the ongoing prediction preferences, and generates fightback soft labels for regularization during training.
A Modular U-Net for Automated Segmentation of X-Ray Tomography Images in Composite Materials
X-Ray Computed Tomography (XCT) techniques have evolved to a point that high-resolution data can be acquired so fast that classic segmentation methods are prohibitively cumbersome, demanding
A Review of Self-supervised Learning Methods in the Field of Medical Image Analysis
  • Jiashu Xu
  • Computer Science
    International Journal of Image, Graphics and Signal Processing
  • 2021
TLDR
This article provides the latest and most detailed overview of self-supervisedLearning in the medical field and promotes the development of unsupervised learning in the field of medical imaging with three categories: context-based, generation- based, and contrast-based.
A modular U-Net for automated segmentation of X-ray tomography images in composite materials
TLDR
A modular interpretation of UNet (Modular U-Net) is proposed and trained to segment 3D tomography images of a three-phased glass fiber-reinforced Polyamide 66 and it is observed that human-comparable results can be achievied even with only 10 annotated layers and using a shallow U- net yields better results than a deeper one.
ABC: Auxiliary Balanced Classifier for Class-imbalanced Semi-supervised Learning
TLDR
This work proposes a scalable class-imbalanced SSL algorithm that can effectively use unlabeled data, while mitigating class imbalance by introducing an auxiliary balanced classifier of a single layer, which is attached to a representation layer of an existing SSL algorithm.
Alleviate Representation Overlapping in Class Incremental Learning by Contrastive Class Concentration
TLDR
A new CIL framework, Contrastive Class Concentration for CIL (C4IL) is proposed to alleviate the phenomenon of representation overlapping for both memorybased and memory-free methods.
An Empirical Investigation of Learning from Biased Toxicity Labels
TLDR
While it is found that initial training on all of the data and fine-tuning on clean data produces models with the highest AUC, it is also found that no single strategy performs best across all fairness metrics.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 58 REFERENCES
Learning Imbalanced Datasets with Label-Distribution-Aware Margin Loss
TLDR
A theoretically-principled label-distribution-aware margin (LDAM) loss motivated by minimizing a margin-based generalization bound is proposed that replaces the standard cross-entropy objective during training and can be applied with prior strategies for training with class-imbalance such as re-weighting or re-sampling.
Learning Deep Representation for Imbalanced Classification
TLDR
The representation learned by this approach, when combined with a simple k-nearest neighbor (kNN) algorithm, shows significant improvements over existing methods on both high- and low-level vision classification tasks that exhibit imbalanced class distribution.
Cost-Sensitive Learning of Deep Feature Representations From Imbalanced Data
TLDR
This paper proposes a cost-sensitive (CoSen) deep neural network, which can automatically learn robust feature representations for both the majority and minority classes, and shows that the proposed approach significantly outperforms the baseline algorithms.
Deep Imbalanced Learning for Face Recognition and Attribute Prediction
TLDR
Cluster-based Large Margin Local Embedding (CLMLE), when combined with a simple k-nearest cluster algorithm, shows significant improvements in accuracy over existing methods on both face recognition and face attribute prediction tasks that exhibit imbalanced class distribution.
Decoupling Representation and Classifier for Long-Tailed Recognition
TLDR
It is shown that it is possible to outperform carefully designed losses, sampling strategies, even complex modules with memory, by using a straightforward approach that decouples representation and classification.
Striking the Right Balance With Uncertainty
TLDR
This paper demonstrates that the Bayesian uncertainty estimates directly correlate with the rarity of classes and the difficulty level of individual samples, and presents a novel framework for uncertainty based class imbalance learning that efficiently utilizes sample and class uncertainty information to learn robust features and more generalizable classifiers.
Imbalanced Deep Learning by Minority Class Incremental Rectification
TLDR
These experimental results demonstrate the performance advantages and model scalability of the proposed batch-wise incremental minority class rectification model over the existing state-of-the-art models for addressing the problem of imbalanced data learning.
Deep Over-sampling Framework for Classifying Imbalanced Data
TLDR
An empirical study using public benchmarks shows that the DOS framework not only counteracts class imbalance better than the existing method, but also improves the performance of the CNN in the standard, balanced settings.
Realistic Evaluation of Deep Semi-Supervised Learning Algorithms
TLDR
This work creates a unified reimplemention and evaluation platform of various widely-used SSL techniques and finds that the performance of simple baselines which do not use unlabeled data is often underreported, that SSL methods differ in sensitivity to the amount of labeled and unlabeling data, and that performance can degrade substantially when the unlabelED dataset contains out-of-class examples.
A systematic study of the class imbalance problem in convolutional neural networks
TLDR
The effect of class imbalance on classification performance is detrimental; the method of addressing class imbalance that emerged as dominant in almost all analyzed scenarios was oversampling; and thresholding should be applied to compensate for prior class probabilities when overall number of properly classified cases is of interest.
...
1
2
3
4
5
...