Confidence estimation in Deep Neural networks via density modelling

TL;DR

This paper investigates the problem of accurately estimating confidence in deep neural networks, proposing a density modelling approach that outperforms softmax in detecting various image distortions and adversarial noise.

Contribution

It introduces a novel confidence measure based on density modelling, addressing the shortcomings of softmax for reliable confidence estimation.

Findings

Density modelling improves confidence estimation in noisy and distorted images.

The proposed measure reduces confidence scores for distorted images, unlike softmax.

Experiments demonstrate better detection of adversarial and non-adversarial distortions.

Abstract

State-of-the-art Deep Neural Networks can be easily fooled into providing incorrect high-confidence predictions for images with small amounts of adversarial noise. Does this expose a flaw with deep neural networks, or do we simply need a better way to estimate confidence? In this paper we consider the problem of accurately estimating predictive confidence. We formulate this problem as that of density modelling, and show how traditional methods such as softmax produce poor estimates. To address this issue, we propose a novel confidence measure based on density modelling approaches. We test these measures on images distorted by blur, JPEG compression, random noise and adversarial noise. Experiments show that our confidence measure consistently shows reduced confidence scores in the presence of such distortions - a property which softmax often lacks.