Physical world assistive signals for deep neural network classifiers -- neither defense nor attack

TL;DR

This paper introduces Assistive Signals, optimized perturbations that enhance deep neural network confidence and accuracy in real-world scenarios, independent of adversarial attack considerations.

Contribution

It proposes a novel concept of assistive signals, extending their optimization to 3D space for real-life conditions, and analyzes their properties and implications.

Findings

Assistive signals improve model confidence and accuracy more than conventional 2D methods.

Assistive signals reveal intrinsic biases of models towards certain object patterns.

Optimization in 3D space enhances robustness under varying lighting and viewing angles.

Abstract

Deep Neural Networks lead the state of the art of computer vision tasks. Despite this, Neural Networks are brittle in that small changes in the input can drastically affect their prediction outcome and confidence. Consequently and naturally, research in this area mainly focus on adversarial attacks and defenses. In this paper, we take an alternative stance and introduce the concept of Assistive Signals, which are optimized to improve a model's confidence score regardless if it's under attack or not. We analyse some interesting properties of these assistive perturbations and extend the idea to optimize assistive signals in the 3D space for real-life scenarios simulating different lighting conditions and viewing angles. Experimental evaluations show that the assistive signals generated by our optimization method increase the accuracy and confidence of deep models more than those generated by conventional methods that work in the 2D space. In addition, our Assistive Signals illustrate the intrinsic bias of ML models towards certain patterns in real-life objects. We discuss how we can exploit these insights to re-think, or avoid, some patterns that might contribute to, or degrade, the detectability of objects in the real-world.