Asymmetric Loss For Multi-Label Classification

Abstract

In a typical multi-label setting, a picture contains on average few positive labels, and many negative ones. This positive-negative imbalance dominates the optimization process, and can lead to under-emphasizing gradients from positive labels during training, resulting in poor accuracy. In this paper, we introduce a novel asymmetric loss ("ASL"), which operates differently on positive and negative samples. The loss enables to dynamically down-weights and hard-thresholds easy negative samples, while also discarding possibly mislabeled samples. We demonstrate how ASL can balance the probabilities of different samples, and how this balancing is translated to better mAP scores. With ASL, we reach state-of-the-art results on multiple popular multi-label datasets: MS-COCO, Pascal-VOC, NUS-WIDE and Open Images. We also demonstrate ASL applicability for other tasks, such as single-label classification and object detection. ASL is effective, easy to implement, and does not increase the training time or complexity. Implementation is available at: https://github.com/Alibaba-MIIL/ASL.

arXiv commentAccepted to ICCV 2021

References in corpus (8)

• Decoupled Weight Decay Regularization
• EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
• Improved Regularization of Convolutional Neural Networks with Cutout
• Focal Loss for Dense Object Detection
• A disciplined approach to neural network hyper-parameters: Part 1 -- learning rate, batch size, momentum, and weight decay
• MMDetection: Open MMLab Detection Toolbox and Benchmark
• The Open Images Dataset V4: Unified image classification, object detection, and visual relationship detection at scale
• Attention-Driven Dynamic Graph Convolutional Network for Multi-Label Image Recognition

Cited by in corpus (17)

• Learning Multi-dimensional Edge Feature-based AU Relation Graph for Facial Action Unit Recognition
• Query2Label: A Simple Transformer Way to Multi-Label Classification
• Sewer-ML: A Multi-Label Sewer Defect Classification Dataset and Benchmark
• Multi-Label Image Classification with Contrastive Learning
• Evaluating Multi-label Classifiers with Noisy Labels
• SCIDA: Self-Correction Integrated Domain Adaptation from Single- to Multi-label Aerial Images
• Re-labeling ImageNet: from Single to Multi-Labels, from Global to Localized Labels
• FOSS: A Self-Learned Doctor for Query Optimizer
• ImageNet-21K Pretraining for the Masses
• Context-Aware Image Inpainting with Learned Semantic Priors
• Hierarchical Knowledge Guided Learning for Real-world Retinal Diseases Recognition
• GraSSNet: Graph Soft Sensing Neural Networks
• Multi-Label Gold Asymmetric Loss Correction with Single-Label Regulators
• GAM: Explainable Visual Similarity and Classification via Gradient Activation Maps
• Overview of Tencent Multi-modal Ads Video Understanding Challenge
• Which Samples Should be Learned First: Easy or Hard?
• Fine-Grained AutoAugmentation for Multi-Label Classification