Detecting Misclassification Errors in Neural Networks with a Gaussian Process Model

As neural network classifiers are deployed in real-world applications, it is crucial that their failures can be detected reliably. One practical solution is to assign confidence scores to each prediction, then use these scores to filter out possible misclassifications. However, existing confidence metrics are not yet sufficiently reliable for this role. This paper presents a new framework that produces a more reliable quantitative metric for detecting misclassification errors. This framework, RED, builds an error detector on top of the base classifier and estimates uncertainty of the detection scores using Gaussian Processes. Empirical comparisons with other error detection methods on 125 UCI datasets demonstrate that this approach is effective. Additional implementations on two probabilistic base classifiers and a large deep learning architecture solving a vision task further confirm the robustness of the method. A case study involving out-of-distribution and adversarial samples shows potential of the proposed method to improve trustworthiness of neural network classifiers more broadly in the future.