AI 中文总结
研究针对深度学习模型在标签噪声下性能下降问题，提出NegScale框架，利用差异不变性，整合结构化负正交惩罚和差异校准相似性调整，经理论分析和实验验证，该框架优于现有基线，建立了新基准。
AI 中文摘要
深度学习模型在视觉识别中表现出色，但训练标签受噪声影响时性能会严重下降。以往工作在标签噪声下无法学习到准确的相似性，误导学习过程。本文发现了一种互补且新颖的现象——差异不变性，即不相关样本间的语义差异在标签噪声下保持稳定。基于此，提出了NegScale框架，它将重点从脆弱的相似性转移到稳健的差异性上。NegScale整合了结构化负正交惩罚和差异校准相似性调整，还进行了理论分析。实验结果表明NegScale优于现有基线，在CIFAR和真实世界数据集上建立了新基准。
英文摘要
Deep learning models excel in visual recognition but suffer severe performance drops when training labels are corrupted by noise. Under label noise prior work cannot learn accurate similarities and thus misguide the learning process. In this paper, we uncover a complementary and novel phenomenon, Dissimilarity Invariance, whereby semantic dissimilarity between unrelated samples remains stable despite label noise. Leveraging this insight, we propose NegScale, a plug-and-play framework that shifts focus from fragile similarity to robust dissimilarity. NegScale integrates: (1) Structured Negative Orthogonality Penalty (SNOP), enforcing subspace orthogonality for unrelated samples; and (2) Dissimilarity-Calibrated Similarity Adjustment (DCSA), suppressing spurious similarity using dissimilarity anchors. We also give theoretical analysis that proves Dissimilarity Invariance and the effectiveness of NegScale. Empirical results demonstrate that NegScale consistently outperforms state-of-the-art baselines, establishing new benchmarks on CIFAR with synthetic noise and real-world datasets.
CommentsAAAI2026
