Logit-based Uncertainty Measure in Classification
This provides a more reliable uncertainty quantification method for classification problems, though it appears incremental as an improvement over existing uncertainty measures.
The paper introduces logit uncertainty, a new uncertainty measure for classification tasks based on neural network logit outputs, and shows it outperforms existing measures on tasks like out-of-sample detection and finding erroneous predictions.
We introduce a new, reliable, and agnostic uncertainty measure for classification tasks called logit uncertainty. It is based on logit outputs of neural networks. We in particular show that this new uncertainty measure yields a superior performance compared to existing uncertainty measures on different tasks, including out of sample detection and finding erroneous predictions. We analyze theoretical foundations of the measure and explore a relationship with high density regions. We also demonstrate how to test uncertainty using intermediate outputs in training of generative adversarial networks. We propose two potential ways to utilize logit-based uncertainty in real world applications, and show that the uncertainty measure outperforms.