Improve Uncertainty Estimation for Unknown Classes in Bayesian Neural Networks with Semi-Supervised /One Set Classification
This addresses uncertainty estimation for unknown classes in BNNs, which is important for safety-critical applications, but the approach appears incremental.
The paper tackles the problem of Bayesian Neural Networks (BNNs) failing to effectively capture uncertainty for unknown classes, which is critical in safety domains like autonomous driving and medical diagnosis. It introduces a simple improvement using one class classification, empirically showing results on MNIST, notMNIST, and FMNIST datasets.
Although deep neural network (DNN) has achieved many state-of-the-art results, estimating the uncertainty presented in the DNN model and the data is a challenging task. Problems related to uncertainty such as classifying unknown classes (class which does not appear in the training data) data as known class with high confidence, is critically concerned in the safety domain area (e.g, autonomous driving, medical diagnosis). In this paper, we show that applying current Bayesian Neural Network (BNN) techniques alone does not effectively capture the uncertainty. To tackle this problem, we introduce a simple way to improve the BNN by using one class classification (in this paper, we use the term "set classification" instead). We empirically show the result of our method on an experiment which involves three datasets: MNIST, notMNIST and FMNIST.