Contextuality Helps Representation Learning for Generalized Category Discovery

TL;DR

This work tackles Generalized Category Discovery (GCD), where unlabeled data contain both known and novel categories. It introduces a contextuality-driven approach with two complementary losses: an instance-level neighborhood contextual loss $\mathcal{L}_n$ and a cluster-level prototypical contextual loss $\mathcal{L}_c$, built on a SimGCD baseline that uses representation learning and a parametric prototype classifier. The method demonstrates improved performance across eight benchmarks, outperforming state-of-the-art GCD methods and showing robustness to real-world data distributions through context-aware learning. The approach advances practical GCD by leveraging multi-level contextual relationships, with code available at the provided repository for reproducibility and further research.

Abstract

This paper introduces a novel approach to Generalized Category Discovery (GCD) by leveraging the concept of contextuality to enhance the identification and classification of categories in unlabeled datasets. Drawing inspiration from human cognition's ability to recognize objects within their context, we propose a dual-context based method. Our model integrates two levels of contextuality: instance-level, where nearest-neighbor contexts are utilized for contrastive learning, and cluster-level, employing prototypical contrastive learning based on category prototypes. The integration of the contextual information effectively improves the feature learning and thereby the classification accuracy of all categories, which better deals with the real-world datasets. Different from the traditional semi-supervised and novel category discovery techniques, our model focuses on a more realistic and challenging scenario where both known and novel categories are present in the unlabeled data. Extensive experimental results on several benchmark data sets demonstrate that the proposed model outperforms the state-of-the-art. Code is available at: https://github.com/Clarence-CV/Contexuality-GCD

Figures (1)

• Figure 1: Two levels of contextuality are explored in our method. The instance-level context leverages the nearest-neighbor context and pseudo-labels to search pair-wise data points for contrastive learning. The cluster-level context forms prototypes to learn the representation via prototypical contrastive learning.