Can I read 2020 International Joint Conference on Neural Networks (IJCNN) - ArcGrad: Angular Gradient Margin Loss for Classification on EtoBox?

2020 International Joint Conference on Neural Networks (IJCNN) - ArcGrad: Angular Gradient Margin Loss for Classification by Wu, Jiantao; Wang, Lin is a scholarly article available to read on EtoBox.

What is 2020 International Joint Conference on Neural Networks (IJCNN) - ArcGrad: Angular Gradient Margin Loss for Classification about?

The cosine-based softmax loss functions greatly enhance intra-class compactness and perform well on the tasks of face recognition and object classification. Outperformance, how ever, depends on the careful hyperparameter selection. Adaptively Scaling Cosine Logits (AdaCos) tries to propose a parameter-free version by leveraging an adaptive scaling parameter. Neverthe less, the application of AdaCos is limited in specific domains because of improper approximation. In this paper, to promote intra-class compactness and inter class separability, we propose an Angular Gradient Margin Loss (ArcGrad) that generates a gradient margin by maximizing the angular gradient. Our work suggests that the margin parameter on cosine-based methods is not necessary, and the scaling pa rameter is inversely proportional to the margin. Furthermore, a stable and large gradient promotes better feature representation. In experiments, we test our method, as well as other methods enhancing discriminative information, on CIFAR and 15 datasets from UCI. Experimental results show ArcGrad consistently outperforms both on large and small scale problems and has the superiority in discriminative information and time-

Author

Wu, Jiantao; Wang, Lin

Publisher

IEEE

Published

2020

Language

EN