Estimating Regression Predictive Distributions with Sample Networks
This work addresses uncertainty estimation for regression tasks in deep learning, which is crucial for real-world applications, representing a novel method for a known bottleneck.
The authors tackled the problem of unreliable uncertainty estimates in deep neural networks due to poor parametric distribution fits by proposing SampleNet, a flexible architecture that learns an empirical distribution using samples, achieving superior performance on large-scale real-world regression tasks.
Estimating the uncertainty in deep neural network predictions is crucial for many real-world applications. A common approach to model uncertainty is to choose a parametric distribution and fit the data to it using maximum likelihood estimation. The chosen parametric form can be a poor fit to the data-generating distribution, resulting in unreliable uncertainty estimates. In this work, we propose SampleNet, a flexible and scalable architecture for modeling uncertainty that avoids specifying a parametric form on the output distribution. SampleNets do so by defining an empirical distribution using samples that are learned with the Energy Score and regularized with the Sinkhorn Divergence. SampleNets are shown to be able to well-fit a wide range of distributions and to outperform baselines on large-scale real-world regression tasks.