Density of States Estimation for Out-of-Distribution Detection
This addresses the issue of poor specificity in probabilistic models for OOD detection, which is crucial for safety in machine learning applications, though it appears incremental as it builds on existing statistical physics concepts.
The paper tackles the problem of out-of-distribution (OOD) detection by proposing DoSE, a density of states estimator that avoids direct model probability comparisons, achieving state-of-the-art performance on established hard benchmarks.
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.