Prediction intervals for Deep Neural Networks
arXiv:2010.04044
Abstract
The aim of this paper is to propose a suitable method for constructing prediction intervals for the output of neural network models. To do this, we adapt the extremely randomized trees method originally developed for random forests to construct ensembles of neural networks. The extra-randomness introduced in the ensemble reduces the variance of the predictions and yields gains in out-of-sample accuracy. An extensive Monte Carlo simulation exercise shows the good performance of this novel method for constructing prediction intervals in terms of coverage probability and mean square prediction error. This approach is superior to state-of-the-art methods extant in the literature such as the widely used MC dropout and bootstrap procedures. The out-of-sample accuracy of the novel algorithm is further evaluated using experimental settings already adopted in the literature.
arXiv comment35 pages, 3 Figures, and 2 Tables
References in corpus (15)
- Deep Learning in Neural Networks: An Overview
- Improving neural networks by preventing co-adaptation of feature detectors
- What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
- Bayesian Convolutional Neural Networks with Bernoulli Approximate Variational Inference
- Probabilistic Backpropagation for Scalable Learning of Bayesian Neural Networks
- A Mean Field View of the Landscape of Two-Layers Neural Networks
- Variational Dropout and the Local Reparameterization Trick
- Deep and Confident Prediction for Time Series at Uber
- Stochastic Gradient Descent Optimizes Over-parameterized Deep ReLU Networks
- Why M Heads are Better than One: Training a Diverse Ensemble of Deep Networks
- High-Quality Prediction Intervals for Deep Learning: A Distribution-Free, Ensembled Approach
- Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing
- An empirical analysis of dropout in piecewise linear networks
- Reliable Prediction Errors for Deep Neural Networks Using Test-Time Dropout
- A Bayesian encourages dropout