☆ 4.7 Article

A survey of class-imbalanced semi-supervised learning

MACHINE LEARNING (2023)

相关参考文献

注意：仅列出部分参考文献，下载原文获取全部文献信息。

Article Computer Science, Artificial Intelligence

DeepSMOTE: Fusing Deep Learning and SMOTE for Imbalanced Data

Damien Dablain et al.

Summary: In this study, we propose a novel oversampling algorithm called DeepSMOTE for deep learning models, which generates high-quality artificial images to increase the number of samples for minority classes and balance the training set.

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2023)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Towards Realistic Semi-supervised Learning

Mamshad Nayeem Rizve et al.

Summary: Deep learning is advancing computer vision applications, but the challenge of capturing real-world data remains. Semi-supervised learning aims to reduce annotation cost by using unlabeled data. This paper introduces the concept of open-world semi-supervised learning and proposes a pseudo-label based approach that incorporates class distribution knowledge. Experimental results show significant improvement over existing state-of-the-art methods on multiple benchmark datasets.

COMPUTER VISION, ECCV 2022, PT XXXI (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

CoSSL: Co-Learning of Representation and Classifier for Imbalanced Semi-Supervised Learning

Yue Fan et al.

Summary: This paper introduces a novel co-learning framework (CoSSL) that addresses the issue of imbalanced data by decoupling representation and classifier learning while closely coupling them. The authors also propose conducting a comprehensive evaluation under various shifted test distributions to improve the practicality of evaluation protocols.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022) (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Masked Autoencoders Are Scalable Vision Learners

Kaiming He et al.

Summary: This paper presents a self-supervised learning method for computer vision based on masked autoencoders. By masking a portion of the input image and reconstructing the missing pixels, large models can be trained efficiently and effectively. The approach achieves high generalization performance and outperforms supervised pretraining in transfer learning tasks.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022) (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Safe-Student for Safe Deep Semi-Supervised Learning with Unseen-Class Unlabeled Data

Rundong He et al.

Summary: This paper proposes a safe SSL method called SAFE-STUDENT that utilizes unlabeled data from the teacher-student view to improve algorithm performance. The method introduces a new scoring function called energy-discrepancy (ED) to enhance instance selection security, and a novel mechanism for learning unseen-class label distribution to mitigate unseen-class perturbation. Experimental results demonstrate that SAFE-STUDENT outperforms existing methods, indicating its feasibility and robustness in addressing the under-explored problem.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022) (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

DASO: Distribution-Aware Semantics-Oriented Pseudo-label for Imbalanced Semi-Supervised Learning

Youngtaek Oh et al.

Summary: This paper introduces a general pseudo-label framework called DASO, which addresses the problem of biased pseudo-labels caused by class imbalance and class distribution mismatch. The framework blends semantic pseudo-labels from a similarity-based classifier with linear pseudo-labels from a linear classifier, and introduces a semantic alignment loss to establish balanced feature representation. Experimental results demonstrate that the DASO framework reliably improves the performance of semi-supervised learning in the presence of both class imbalance and distribution mismatch.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Influence-Balanced Loss for Imbalanced Visual Classification

Seulki Park et al.

Summary: This paper proposes a balancing training method with a new loss function that improves the performance of imbalanced data learning. Experimental results demonstrate the effectiveness of the proposed method and its superiority over state-of-the-art cost-sensitive loss methods.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Adversarial Robustness under Long-Tailed Distribution

Tong Wu et al.

Summary: This study investigates adversarial vulnerability and defense mechanisms under long-tailed distributions, revealing the negative impacts of imbalanced data and proposing a new framework RoBal with two dedicated modules to enhance adversarial robustness. Experiment results show the superiority of the proposed approach over state-of-the-art defense methods, making it a significant step towards real-world robustness.

2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021 (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Deep Imbalanced Learning for Face Recognition and Attribute Prediction

Chen Huang et al.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2020)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learning

Takeru Miyato et al.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2019)

添加到收藏夹

Review Multidisciplinary Sciences

A brief introduction to weakly supervised learning

Zhi-Hua Zhou

NATIONAL SCIENCE REVIEW (2018)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Multiple instance learning: A survey of problem characteristics and applications

Marc-Andre Carbonneau et al.

PATTERN RECOGNITION (2018)

添加到收藏夹

Article Computer Science, Artificial Intelligence

A systematic study of the class imbalance problem in convolutional neural networks

Mateusz Buda et al.

NEURAL NETWORKS (2018)

添加到收藏夹

Proceedings Paper Computer Science, Theory & Methods

Dynamic Sampling in Convolutional Neural Networks for Imbalanced Data Classification

Samira Pouyanfar et al.

IEEE 1ST CONFERENCE ON MULTIMEDIA INFORMATION PROCESSING AND RETRIEVAL (MIPR 2018) (2018)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Redundancy-driven modified Tomek-link based undersampling: A solution to class imbalance

Debashree Devi et al.

PATTERN RECOGNITION LETTERS (2017)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Relay Backpropagation for Effective Learning of Deep Convolutional Neural Networks

Li Shen et al.

COMPUTER VISION - ECCV 2016, PT VII (2016)

添加到收藏夹

Article Chemistry, Analytical

Semi-Supervised Bayesian Classification of Materials with Impact-Echo Signals

Jorge Igual et al.

SENSORS (2015)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Classification in the Presence of Label Noise: a Survey

Benoit Frenay et al.

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2014)

添加到收藏夹

Review Computer Science, Artificial Intelligence

A review of multi-instance learning assumptions

James Foulds et al.

KNOWLEDGE ENGINEERING REVIEW (2010)

添加到收藏夹

Review Computer Science, Artificial Intelligence

Learning from Imbalanced Data

Haibo He et al.

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING (2009)

添加到收藏夹

Article Automation & Control Systems

Exploratory Undersampling for Class-Imbalance Learning

Xu-Ying Liu et al.

IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART B-CYBERNETICS (2009)

添加到收藏夹

© Peeref 2019-2024. All rights reserved.