Self-Cooperation Knowledge Distillation for Novel Class Discovery

TL;DR

This paper introduces Self-Cooperation Knowledge Distillation, a novel approach for Novel Class Discovery that addresses class imbalance by leveraging self-cooperation between known and unknown class representations, leading to state-of-the-art results.

Contribution

The paper proposes a self-cooperation knowledge distillation method that constructs disjoint representation spaces for known and novel classes and uses mutual information to improve discovery performance.

Findings

Achieves significant performance improvements on six datasets.

Outperforms existing methods, reaching state-of-the-art results.

Effectively handles class imbalance in NCD tasks.

Abstract

Novel Class Discovery (NCD) aims to discover unknown and novel classes in an unlabeled set by leveraging knowledge already learned about known classes. Existing works focus on instance-level or class-level knowledge representation and build a shared representation space to achieve performance improvements. However, a long-neglected issue is the potential imbalanced number of samples from known and novel classes, pushing the model towards dominant classes. Therefore, these methods suffer from a challenging trade-off between reviewing known classes and discovering novel classes. Based on this observation, we propose a Self-Cooperation Knowledge Distillation (SCKD) method to utilize each training sample (whether known or novel, labeled or unlabeled) for both review and discovery. Specifically, the model's feature representations of known and novel classes are used to construct two disjoint representation spaces. Through spatial mutual information, we design a self-cooperation learning to encourage model learning from the two feature representation spaces from itself. Extensive experiments on six datasets demonstrate that our method can achieve significant performance improvements, achieving state-of-the-art performance.