Quantification of Predictive Uncertainty via Inference-Time Sampling
This addresses uncertainty quantification for users of feed-forward networks, offering a flexible, architecture-agnostic solution, though it is incremental as it builds on existing sampling methods.
The paper tackles the problem of predictive uncertainty in deterministic neural networks by proposing a post-hoc sampling strategy that generates diverse outputs without architectural changes, showing a correlation between estimated uncertainty and prediction error in regression tasks.
Predictive variability due to data ambiguities has typically been addressed via construction of dedicated models with built-in probabilistic capabilities that are trained to predict uncertainty estimates as variables of interest. These approaches require distinct architectural components and training mechanisms, may include restrictive assumptions and exhibit overconfidence, i.e., high confidence in imprecise predictions. In this work, we propose a post-hoc sampling strategy for estimating predictive uncertainty accounting for data ambiguity. The method can generate different plausible outputs for a given input and does not assume parametric forms of predictive distributions. It is architecture agnostic and can be applied to any feed-forward deterministic network without changes to the architecture or training procedure. Experiments on regression tasks on imaging and non-imaging input data show the method's ability to generate diverse and multi-modal predictive distributions, and a desirable correlation of the estimated uncertainty with the prediction error.