NUQ: Nonparametric Uncertainty Quantification for Deterministic Neural Networks

02/07/2022
by   Nikita Kotelevskii, et al.
0

This paper proposes a fast and scalable method for uncertainty quantification of machine learning models' predictions. First, we show the principled way to measure the uncertainty of predictions for a classifier based on Nadaraya-Watson's nonparametric estimate of the conditional label distribution. Importantly, the approach allows to disentangle explicitly aleatoric and epistemic uncertainties. The resulting method works directly in the feature space. However, one can apply it to any neural network by considering an embedding of the data induced by the network. We demonstrate the strong performance of the method in uncertainty estimation tasks on a variety of real-world image datasets, such as MNIST, SVHN, CIFAR-100 and several versions of ImageNet.

READ FULL TEXT
research
02/05/2023

Direct Uncertainty Quantification

Traditional neural networks are simple to train but they produce overcon...
research
08/17/2023

Discretization-Induced Dirichlet Posterior for Robust Uncertainty Quantification on Regression

Uncertainty quantification is critical for deploying deep neural network...
research
04/01/2022

DBCal: Density Based Calibration of classifier predictions for uncertainty quantification

Measurement of uncertainty of predictions from machine learning methods ...
research
09/17/2022

Sample-based Uncertainty Quantification with a Single Deterministic Neural Network

Development of an accurate, flexible, and numerically efficient uncertai...
research
03/04/2020

Simple and Scalable Epistemic Uncertainty Estimation Using a Single Deep Deterministic Neural Network

We propose a method for training a deterministic deep model that can fin...
research
12/28/2022

A Learning-Based Optimal Uncertainty Quantification Method and Its Application to Ballistic Impact Problems

This paper concerns the study of optimal (supremum and infimum) uncertai...
research
02/21/2023

Improved uncertainty quantification for neural networks with Bayesian last layer

Uncertainty quantification is an essential task in machine learning - a ...

Please sign up or login with your details

Forgot password? Click here to reset