arXiv cs.AI· Zitai Wang, Qianqian Xu, Cong Hua, Xi Chen, Zhiyong Yang, Xiaochun Cao, Qingming Huang·· 14 小时前AI 评分18
广义类别发现中的语义错映射：评估与缓解
Semantic Mismapping in Generalized Category Discovery: Evaluation and Mitigation
AI 导读
研究指出，广义类别发现（GCD）方法沿用全局匈牙利匹配会重新分配已锚定的旧类索引，导致新类准确率被高估最多 25.63 点，作者将这一现象称为语义错映射，并给出置换环条件刻画其何时抬高匹配分数。
正文 · 原文
Abstract:Generalized Category Discovery (GCD) aims to perform old-category recognition and novel-category discovery in the same unlabeled set. Early methods address both objectives through clustering and evaluate the exchangeable predicted indices with global Hungarian (GH) matching. Recent methods adopt trainable parametric heads, using supervision for old-category recognition and discovery-oriented objectives for new categories. In this way, the output space becomes partially exchangeable: old indices are semantically anchored, while new indices remain exchangeable. However, these methods still apply global matching, which can reassign anchored old indices to other categories and hide some prediction errors. Empirically, we are surprised to observe that this reassignment can induce an overestimation of new-category accuracy by up to 25.63 points. We term this phenomenon semantic mismapping and further derive a permutation-cycle condition that characterizes when semantic mismapping can increase the global matching score. To fix this issue, we propose Semantics-Preserving Evaluation (SPE), which preserves the supervised semantics of old indices and restricts Hungarian matching to the new region. Furthermore, we propose Mismapping-Mitigating Optimization (MMO) to alleviate the prediction errors that can give rise to semantic mismapping, combining a region-balance loss (RBL) for old/new routing with a novel-diversity loss (NDL) that encourages orthogonality among new prototypes. Theoretically, we show that minimizing the two losses can tighten an upper bound on spurious accuracy gains from semantic mismapping, with controllable residual terms. Empirical results under both GH and SPE demonstrate the effectiveness of our proposed method.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.39356 [cs.CV] |
| (or arXiv:2610.39356v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.39356 arXiv-issued DOI via DataCite |
Submission history
From: Zitai Wang [view email]
[v1]
Thu, 29 Oct 2026 04:49:17 UTC (15,809 KB)
来源：arXiv cs.AI · arxiv.org