A Framework for Uncertainty Quantification Based on Nearest Neighbors Across Layers

TL;DR

This paper introduces a post-hoc uncertainty quantification framework for neural networks using nearest neighbors across layers, improving detection of uncertain predictions especially in challenging classification tasks.

Contribution

It proposes a novel method leveraging nearest neighbors and new metrics to better estimate uncertainty in neural network decisions after training.

Findings

Metrics improve uncertainty estimation over softmax confidence.

Method outperforms existing confidence measures on CIFAR-10 and MNIST.

Enhances reliability in high-risk applications like medical diagnosis.

Abstract

Neural Networks have high accuracy in solving problems where it is difficult to detect patterns or create a logical model. However, these algorithms sometimes return wrong solutions, which become problematic in high-risk domains like medical diagnosis or autonomous driving. One strategy to detect and mitigate these errors is the measurement of the uncertainty over neural network decisions. In this paper, we present a novel post-hoc framework for measuring the uncertainty of a decision based on retrieved training cases that have a similar activation vector to the query for each layer. Based on these retrieved cases, we propose two new metrics: Decision Change and Layer Uncertainty, which capture changes in nearest-neighbor class distributions across layers. We evaluated our approach in a classification model for two datasets: CIFAR-10 and MNIST. The results show that these metrics enhance uncertainty estimation, especially in challenging classification tasks, outperforming softmax-based confidence.