Improving Uncertainty Calibration via Prior Augmented Data
This addresses the issue of unreliable probabilistic predictions in neural networks for users in safety-critical applications, though it is an incremental improvement over existing calibration methods.
The paper tackles the problem of neural networks being overconfident and miscalibrated, especially when test data differs from training data, by adjusting predictions to increase entropy in overconfident regions, resulting in improved calibration across classification and regression tasks.
Neural networks have proven successful at learning from complex data distributions by acting as universal function approximators. However, they are often overconfident in their predictions, which leads to inaccurate and miscalibrated probabilistic predictions. The problem of overconfidence becomes especially apparent in cases where the test-time data distribution differs from that which was seen during training. We propose a solution to this problem by seeking out regions of feature space where the model is unjustifiably overconfident, and conditionally raising the entropy of those predictions towards that of the prior distribution of the labels. Our method results in a better calibrated network and is agnostic to the underlying model structure, so it can be applied to any neural network which produces a probability density as an output. We demonstrate the effectiveness of our method and validate its performance on both classification and regression problems, applying it to recent probabilistic neural network models.