对比四种训练目标在图像分布外检测中的表现，发现交叉熵损失最稳定。
A Systematic Comparison of Training Objectives for Out-of-Distribution Detection in Image Classification
- 系统比较交叉熵、原型、三元组和平均精度四类训练目标
- 交叉熵损失在近域和远域分布外检测上均保持最高且稳定的AUROC
- 其他方法在特定场景下表现可媲美，适合针对性优化
分布外（OOD）检测在安全敏感应用中至关重要。尽管该问题已从多个角度被研究，但训练目标对OOD性能的影响仍相对未被充分探索。本文系统比较了四种广泛使用的训练目标：交叉熵损失、原型损失、三元组损失和平均精度（AP）损失，涵盖概率性、基于原型、度量学习和排序监督，评估其在图像分类中基于标准化OpenOOD协议的分布外检测能力。在所采用的ResNet-18/OpenOOD设置及各目标对应的OOD评分规则下，交叉熵损失、原型损失与AP损失在分布内准确率上表现相当，而交叉熵损失在近域与远域分布外检测的整体AUROC上最为一致且最优；其余目标在特定条件下亦具竞争力。
原文摘要 · Abstract (English)
Out-of-distribution (OOD) detection is critical in safety-sensitive applications. While this challenge has been addressed from various perspectives, the influence of training objectives on OOD behavior remains comparatively underexplored. In this paper, we present a systematic comparison of four widely used training objectives: Cross-Entropy Loss, Prototype Loss, Triplet Loss, and Average Precision (AP) Loss, spanning probabilistic, prototype-based, metric-learning, and ranking-based supervision, for OOD detection in image classification under standardized OpenOOD protocols. Within the evaluated ResNet-18/OpenOOD setting and objective-specific OOD scoring rules, Cross-Entropy Loss, Prototype Loss, and AP Loss achieve comparable in-distribution accuracy, while Cross-Entropy Loss provides the most consistent near- and far-OOD AUROC overall; the other objectives can be competitive in specific settings.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。