Robust Classification with Convolutional Prototype Learning

TL;DR

This paper introduces convolutional prototype learning (CPL), a new framework that enhances CNN robustness against adversarial examples and open world recognition by using prototypes and a prototype loss for better feature representation.

Contribution

The paper proposes CPL, a novel learning framework that improves CNN robustness and open world recognition by incorporating prototypes and a regularization loss.

Findings

CPL achieves comparable or better accuracy than traditional CNNs.

CPL demonstrates superior robustness in rejection and incremental learning tasks.

Experiments validate the effectiveness of prototypes and prototype loss in enhancing model robustness.

Abstract

Convolutional neural networks (CNNs) have been widely used for image classification. Despite its high accuracies, CNN has been shown to be easily fooled by some adversarial examples, indicating that CNN is not robust enough for pattern classification. In this paper, we argue that the lack of robustness for CNN is caused by the softmax layer, which is a totally discriminative model and based on the assumption of closed world (i.e., with a fixed number of categories). To improve the robustness, we propose a novel learning framework called convolutional prototype learning (CPL). The advantage of using prototypes is that it can well handle the open world recognition problem and therefore improve the robustness. Under the framework of CPL, we design multiple classification criteria to train the network. Moreover, a prototype loss (PL) is proposed as a regularization to improve the intra-class compactness of the feature representation, which can be viewed as a generative model based on the Gaussian assumption of different classes. Experiments on several datasets demonstrate that CPL can achieve comparable or even better results than traditional CNN, and from the robustness perspective, CPL shows great advantages for both the rejection and incremental category learning tasks.