Exploring the Tradeoff Between Diversity and Discrimination for Continuous Category Discovery

TL;DR

This paper introduces IDOD, a novel method for continuous category discovery that enhances diversity, reduces error accumulation, and mitigates forgetting with lower storage needs, outperforming existing approaches on fine-grained datasets.

Contribution

The paper proposes IDOD, a new framework combining independent enrichment, joint discovery, and orthogonality modules to improve continuous category discovery.

Findings

IDOD outperforms state-of-the-art methods on fine-grained datasets.

It effectively balances diversity and discrimination in category discovery.

The method reduces storage overhead while preventing catastrophic forgetting.

Abstract

Continuous category discovery (CCD) aims to automatically discover novel categories in continuously arriving unlabeled data. This is a challenging problem considering that there is no number of categories and labels in the newly arrived data, while also needing to mitigate catastrophic forgetting. Most CCD methods cannot handle the contradiction between novel class discovery and classification well. They are also prone to accumulate errors in the process of gradually discovering novel classes. Moreover, most of them use knowledge distillation and data replay to prevent forgetting, occupying more storage space. To address these limitations, we propose Independence-based Diversity and Orthogonality-based Discrimination (IDOD). IDOD mainly includes independent enrichment of diversity module, joint discovery of novelty module, and continuous increment by orthogonality module. In independent enrichment, the backbone is trained separately using contrastive loss to avoid it focusing only on features for classification. Joint discovery transforms multi-stage novel class discovery into single-stage, reducing error accumulation impact. Continuous increment by orthogonality module generates mutually orthogonal prototypes for classification and prevents forgetting with lower space overhead via representative representation replay. Experimental results show that on challenging fine-grained datasets, our method outperforms the state-of-the-art methods.