Neighborhood Commonality-aware Evolution Network for Continuous Generalized Category Discovery

TL;DR

This paper introduces NCENet, a novel framework for continuous generalized category discovery that leverages neighborhood commonalities and contrastive knowledge distillation to improve representation learning for old and new classes.

Contribution

The paper proposes NCENet, combining neighborhood commonality-aware representation learning and bi-level contrastive knowledge distillation for improved continual class discovery.

Findings

Outperforms previous state-of-the-art methods on CIFAR10, CIFAR100, and Tiny-ImageNet.

Achieves 3.09% higher clustering accuracy on old classes in CIFAR100.

Achieves 6.32% higher clustering accuracy on new classes in CIFAR100.

Abstract

Continuous Generalized Category Discovery (C-GCD) aims to continually discover novel classes from unlabelled image sets while maintaining performance on old classes. In this paper, we propose a novel learning framework, dubbed Neighborhood Commonality-aware Evolution Network (NCENet) that conquers this task from the perspective of representation learning. Concretely, to learn discriminative representations for novel classes, a Neighborhood Commonality-aware Representation Learning (NCRL) is designed, which exploits local commonalities derived neighborhoods to guide the learning of representational differences between instances of different classes. To maintain the representation ability for old classes, a Bi-level Contrastive Knowledge Distillation (BCKD) module is designed, which leverages contrastive learning to perceive the learning and learned knowledge and conducts knowledge distillation. Extensive experiments conducted on CIFAR10, CIFAR100, and Tiny-ImageNet demonstrate the superior performance of NCENet compared to the previous state-of-the-art method. Particularly, in the last incremental learning session on CIFAR100, the clustering accuracy of NCENet outperforms the second-best method by a margin of 3.09\% on old classes and by a margin of 6.32\% on new classes. Our code will be publicly available at \href{https://github.com/xjtuYW/NCENet.git}{https://github.com/xjtuYW/NCENet.git}. \end{abstract}