paper-with-me

홈 › Papers

Imbalance Robust Softmax for Deep Embeeding Learning

2020-11-23 · Hao Zhu, Yang Yuan, Guosheng Hu, Xiang Wu, Neil Robertson

Deep embedding learning is expected to learn a metric space in which features have smaller maximal intra-class distance than minimal inter-class distance. In recent years, one research focus is to solve the open-set problem by discriminative deep embedding learning in the field of face recognition (FR) and person re-identification (re-ID). Apart from open-set problem, we find that imbalanced training data is another main factor causing the performance degradation of FR and re-ID, and data imbalance widely exists in the real applications. However, very little research explores why and how data imbalance influences the performance of FR and re-ID with softmax or its variants. In this work, we deeply investigate data imbalance in the perspective of neural network optimisation and feature distribution about softmax. We find one main reason of performance degradation caused by data imbalance is that the weights (from the penultimate fully-connected layer) are far from their class centers in feature space. Based on this investigation, we propose a unified framework, Imbalance-Robust Softmax (IR-Softmax), which can simultaneously solve the open-set problem and reduce the influence of data imbalance. IR-Softmax can generalise to any softmax and its variants (which are discriminative for open-set problem) by directly setting the weights as their class centers, naturally solving the data imbalance problem. In this work, we explicitly re-formulate two discriminative softmax (A-Softmax and AM-Softmax) under the framework of IR-Softmax. We conduct extensive experiments on FR databases (LFW, MegaFace) and re-ID database (Market-1501, Duke), and IR-Softmax outperforms many state-of-the-art methods.

📄 PDF Abstract BibTeX arXiv:2011.11155

Code (0)

등록된 구현이 없습니다.

Tasks

Face RecognitionPerson Re-Identification

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Similar Papers 제목 키워드 기반

Enlarged Large Margin Loss for Imbalanced Classification

2023-06-15 · Sota Kato, Kazuhiro Hotta

We propose a novel loss function for imbalanced classification. LDAM loss, which minimizes a margin-based generalization bound, is widely utilized for class-imbalanced image classification. Although, by using LDAM loss, …

Classificationimage-classificationImage Classificationimbalanced classification

Neural Network Classifier as Mutual Information Evaluator

2021-06-19 · Zhenyue Qin, Dongwoo Kim, Tom Gedeon

Cross-entropy loss with softmax output is a standard choice to train neural network classifiers. We give a new view of neural network classifiers with softmax and cross-entropy as mutual information evaluators. We show t…

Form

Differentiable Hebbian Plasticity for Continual Learning

2019-05-16 · ICML Workshop AMTL 2019 6 · Anonymous

Catastrophic forgetting poses a grand challenge for continual learning systems, which prevents neural networks from protecting old knowledge while learning new tasks sequentially. We propose a Differentiable Hebbian Plas…

Continual LearningPermuted-MNISTSplit-MNIST

Online Continual Learning via Logit Adjusted Softmax

2023-11-11 · Zhehao Huang, Tao Li, Chenhe Yuan, Yingwen Wu 외

Online continual learning is a challenging problem where models must learn from a non-stationary data stream while avoiding catastrophic forgetting. Inter-class imbalance during training has been identified as a major ca…

Continual Learning

Imbalanced Malware Images Classification: a CNN based Approach

2017-08-27 · Songqing Yue, Tianyang Wang

Deep convolutional neural networks (CNNs) can be applied to malware binary detection via image classification. The performance, however, is degraded due to the imbalance of malware families (classes). To mitigate this is…

ClassificationGeneral Classificationimage-classificationImage Classification