Out-of-distribution (OOD) detection plays a crucial role in ensuring the security of neural networks. Existing works have leveraged the fact that In-distribution (ID) samples form a subspace in the feature space, achieving state-of-the-art (SOTA) performance. However, the comprehensive characteristics of the ID subspace still leave under-explored. Recently, the discovery of Neural Collapse ($\mathcal{NC}$) sheds light on novel properties of the ID subspace. Leveraging insight from $\mathcal{NC}$, we observe that the Principal Angle between the features and the ID feature subspace forms a superior representation for measuring the likelihood of OOD. Building upon this observation, we propose a novel $\mathcal{NC}$-inspired OOD scoring function, named Entropy-enhanced Principal Angle (EPA), which integrates both the global characteristic of the ID subspace and its inner property. We experimentally compare EPA with various SOTA approaches, validating its superior performance and robustness across different network architectures and OOD datasets.

翻译：离群检测在确保神经网络安全性方面发挥着关键作用。现有研究利用同分布样本在特征空间中构成子空间这一特性，取得了最先进的性能。然而，同分布子空间的全面特征仍待深入探索。近期，神经坍缩（$\mathcal{NC}$）的发现揭示了同分布子空间的新性质。借助$\mathcal{NC}$的洞察，我们观察到特征与同分布特征子空间之间的主角能够形成衡量离群可能性的优越表征。基于这一发现，我们提出了一种受$\mathcal{NC}$启发的离群评分函数——熵增强主角（EPA），该函数融合了同分布子空间的全局特征及其内在属性。我们通过实验将EPA与多种最先进方法进行对比，验证了其在不同网络架构和离群数据集上的卓越性能与鲁棒性。