Abstract
Deep learning methods continue to have a decided impact on machine learning, both in theory and in practice. Statistical theoretical developments have been mostly concerned with approximability or rates of estimation when recovering infinite dimensional objects (curves or densities). Despite the impressive array of available theoretical results, the literature has been largely silent about uncertainty quantification for deep learning. This paper takes a step forward in this important direction by taking a Bayesian point of view. We study Gaussian approximability of certain aspects of posterior distributions of sparse deep ReLU architectures in non-parametric regression. Building on tools from Bayesian non-parametrics, we provide semi-parametric Bernstein-von Mises theorems for linear and quadratic functionals, which guarantee that implied Bayesian credible regions have valid frequentist coverage. Our results provide new theoretical justifications for (Bayesian) deep learning with ReLU activation functions, highlighting their inferential potential.
Original language | English (US) |
---|---|
Pages (from-to) | 298-308 |
Number of pages | 11 |
Journal | Proceedings of Machine Learning Research |
Volume | 108 |
State | Published - 2020 |
Externally published | Yes |
Event | 23rd International Conference on Artificial Intelligence and Statistics, AISTATS 2020 - Virtual, Online Duration: Aug 26 2020 → Aug 28 2020 |
ASJC Scopus subject areas
- Artificial Intelligence
- Software
- Control and Systems Engineering
- Statistics and Probability