论文标题
状态估计的密度分布检测
Density of States Estimation for Out-of-Distribution Detection
论文作者
论文摘要
也许令人惊讶的是，最近的研究表明，概率模型可能性的特异性较差（OOD）检测，并且通常将更高的可能性分配给OOD数据，而不是分布数据。为了改善这个问题，我们提出了剂量，即估计量的密度。借助'状态密度'的统计物理概念，剂量决策规则避免了直接比较模型概率，而是利用``模型概率的概率''，或实际上任何合理统计量的频率。使用非参数密度估计器（例如KDE和One Class SVM）计算该频率，该估计值测量了训练数据的各种模型统计数据的典型性，并且我们可以从中以低典型性为异常的测试点标记。与许多其他方法不同，剂量既不需要标记的数据也不需要OOD示例。剂量是模块化的，可以琐碎地应用于任何现有的训练有素的模型。我们在先前确定的``硬''基准上展示了剂量对其他无监督的OOD探测器的最先进的表现。
Perhaps surprisingly, recent studies have shown probabilistic model likelihoods have poor specificity for out-of-distribution (OOD) detection and often assign higher likelihoods to OOD data than in-distribution data. To ameliorate this issue we propose DoSE, the density of states estimator. Drawing on the statistical physics notion of ``density of states,'' the DoSE decision rule avoids direct comparison of model probabilities, and instead utilizes the ``probability of the model probability,'' or indeed the frequency of any reasonable statistic. The frequency is calculated using nonparametric density estimators (e.g., KDE and one-class SVM) which measure the typicality of various model statistics given the training data and from which we can flag test points with low typicality as anomalous. Unlike many other methods, DoSE requires neither labeled data nor OOD examples. DoSE is modular and can be trivially applied to any existing, trained model. We demonstrate DoSE's state-of-the-art performance against other unsupervised OOD detectors on previously established ``hard'' benchmarks.
