ProtoGCD：面向广义类别发现的统一且无偏的原型学习框架

ProtoGCD: Unified and Unbiased Prototype Learning for Generalized Category Discovery

April 2, 2025

作者: Shijie Ma, Fei Zhu, Xu-Yao Zhang, Cheng-Lin Liu

cs.AI

摘要

廣義類別發現（GCD）是一個實用但尚未充分探索的問題，它要求模型通過利用舊類別的標記樣本來自動聚類並發現新類別。其挑戰在於未標記數據中同時包含舊類別和新類別。早期工作利用參數化分類器進行偽標籤處理，將舊類別和新類別分開處理，導致兩者之間的準確性不平衡。最近的方法採用對比學習，卻忽略了潛在的正樣本，並與聚類目標脫鉤，導致表示偏見和次優結果。為解決這些問題，我們引入了一個統一且無偏的原型學習框架，即ProtoGCD，其中舊類別和新類別通過聯合原型和統一學習目標進行建模，實現了舊類別與新類別的統一建模。具體而言，我們提出了一種雙層自適應偽標籤機制來緩解確認偏誤，並結合兩個正則化項共同幫助學習更適合GCD的表示。此外，出於實際考慮，我們設計了一個標準來估計新類別的數量。進一步地，我們將ProtoGCD擴展到檢測未見的異常值，實現了任務層面的統一。綜合實驗表明，ProtoGCD在通用和細粒度數據集上均達到了最先進的性能。代碼可在https://github.com/mashijie1028/ProtoGCD獲取。

English

Generalized category discovery (GCD) is a pragmatic but underexplored problem, which requires models to automatically cluster and discover novel categories by leveraging the labeled samples from old classes. The challenge is that unlabeled data contain both old and new classes. Early works leveraging pseudo-labeling with parametric classifiers handle old and new classes separately, which brings about imbalanced accuracy between them. Recent methods employing contrastive learning neglect potential positives and are decoupled from the clustering objective, leading to biased representations and sub-optimal results. To address these issues, we introduce a unified and unbiased prototype learning framework, namely ProtoGCD, wherein old and new classes are modeled with joint prototypes and unified learning objectives, {enabling unified modeling between old and new classes}. Specifically, we propose a dual-level adaptive pseudo-labeling mechanism to mitigate confirmation bias, together with two regularization terms to collectively help learn more suitable representations for GCD. Moreover, for practical considerations, we devise a criterion to estimate the number of new classes. Furthermore, we extend ProtoGCD to detect unseen outliers, achieving task-level unification. Comprehensive experiments show that ProtoGCD achieves state-of-the-art performance on both generic and fine-grained datasets. The code is available at https://github.com/mashijie1028/ProtoGCD.