Uncertainty Estimation Using a Single Deep Deterministic Neural Network
This addresses the problem of reliable uncertainty quantification for deep learning practitioners, offering a more efficient alternative to ensemble methods.
The paper tackles uncertainty estimation in deep learning by proposing a deterministic neural network method (DUQ) that can identify and reject out-of-distribution data with a single forward pass, matching or improving upon Deep Ensembles on challenging dataset pairs like FashionMNIST vs. MNIST.
We propose a method for training a deterministic deep model that can find and reject out of distribution data points at test time with a single forward pass. Our approach, deterministic uncertainty quantification (DUQ), builds upon ideas of RBF networks. We scale training in these with a novel loss function and centroid updating scheme and match the accuracy of softmax models. By enforcing detectability of changes in the input using a gradient penalty, we are able to reliably detect out of distribution data. Our uncertainty quantification scales well to large datasets, and using a single model, we improve upon or match Deep Ensembles in out of distribution detection on notable difficult dataset pairs such as FashionMNIST vs. MNIST, and CIFAR-10 vs. SVHN.