Relational Retrieval: Leveraging Known-Novel Interactions for Generalized Category Discovery

TL;DR

This paper introduces Relational Pattern Consistency (RPC), a novel method for Generalized Category Discovery that enhances mutual learning between labeled and unlabeled data through relational pattern matching.

Contribution

It proposes RPC, a bidirectional approach that leverages known and unknown interactions for improved category discovery and state-of-the-art results.

Findings

RPC achieves superior performance on benchmark datasets.

Relational pattern matching improves pseudo-label reliability.

Bidirectional knowledge transfer enhances category discovery.

Abstract

In this study, we tackle Generalized Category Discovery (GCD) via a Relational Retrieval perspective, explicitly coupling labeled and unlabeled data through bidirectional knowledge transfer. While existing methods treat these sources separately, missing valuable interaction opportunities, we propose Relational Pattern Consistency (RPC) that enables mutual enhancement. RPC employs One-vs-All classifiers for soft ID/OOD decomposition, then introduces two mechanisms: (i) for known-class preservation, we transfer semantic behavioral alignment; (ii) for category discovery, we leverage the insight that samples from the same category maintain invariant relationships with known-class prototypes, transforming unreliable pseudo-labeling into well-defined relational pattern matching. This bidirectional design allows labeled data to guide unlabeled learning while discovering novel categories through their collective relational signatures. Extensive experiments demonstrate RPC achieves state-of-the-art performance on both generic and fine-grained benchmarks.