Unreliable Uncertainty Estimates with Monte Carlo Dropout

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Djupskås, Aslak, Stasik, Alexander Johannes, Riemer-Sørensen, Signe
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917150758797312
author Djupskås, Aslak
Stasik, Alexander Johannes
Riemer-Sørensen, Signe
author_facet Djupskås, Aslak
Stasik, Alexander Johannes
Riemer-Sørensen, Signe
contents Reliable uncertainty estimation is crucial for machine learning models, especially in safety-critical domains. While exact Bayesian inference offers a principled approach, it is often computationally infeasible for deep neural networks. Monte Carlo dropout (MCD) was proposed as an efficient approximation to Bayesian inference in deep learning by applying neuron dropout at inference time \citep{gal2016dropout}. Hence, the method generates multiple sub-models yielding a distribution of predictions to estimate uncertainty. We empirically investigate its ability to capture true uncertainty and compare to Gaussian Processes (GP) and Bayesian Neural Networks (BNN). We find that MCD struggles to accurately reflect the underlying true uncertainty, particularly failing to capture increased uncertainty in extrapolation and interpolation regions as observed in Bayesian models. The findings suggest that uncertainty estimates from MCD, as implemented and evaluated in these experiments, is not as reliable as those from traditional Bayesian approaches for capturing epistemic and aleatoric uncertainty.
format Preprint
id arxiv_https___arxiv_org_abs_2512_14851
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Unreliable Uncertainty Estimates with Monte Carlo Dropout
Djupskås, Aslak
Stasik, Alexander Johannes
Riemer-Sørensen, Signe
Machine Learning
Reliable uncertainty estimation is crucial for machine learning models, especially in safety-critical domains. While exact Bayesian inference offers a principled approach, it is often computationally infeasible for deep neural networks. Monte Carlo dropout (MCD) was proposed as an efficient approximation to Bayesian inference in deep learning by applying neuron dropout at inference time \citep{gal2016dropout}. Hence, the method generates multiple sub-models yielding a distribution of predictions to estimate uncertainty. We empirically investigate its ability to capture true uncertainty and compare to Gaussian Processes (GP) and Bayesian Neural Networks (BNN). We find that MCD struggles to accurately reflect the underlying true uncertainty, particularly failing to capture increased uncertainty in extrapolation and interpolation regions as observed in Bayesian models. The findings suggest that uncertainty estimates from MCD, as implemented and evaluated in these experiments, is not as reliable as those from traditional Bayesian approaches for capturing epistemic and aleatoric uncertainty.
title Unreliable Uncertainty Estimates with Monte Carlo Dropout
topic Machine Learning
url https://arxiv.org/abs/2512.14851