Physical world assistive signals for deep neural network classifiers -- neither defense nor attack
arXiv:2105.00622
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.
References in corpus (15)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- Towards Deep Learning Models Resistant to Adversarial Attacks
- Random Erasing Data Augmentation
- Towards Deep Neural Network Architectures Robust to Adversarial Examples
- Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey
- DPatch: An Adversarial Patch Attack on Object Detectors
- Soft Rasterizer: A Differentiable Renderer for Image-based 3D Reasoning
- Accelerating 3D Deep Learning with PyTorch3D
- Certified Defenses for Adversarial Patches
- Kaolin: A PyTorch Library for Accelerating 3D Deep Learning Research
- Enhancing Adversarial Example Transferability with an Intermediate Level Attack
- Exploring the Vulnerability of Single Shot Module in Object Detectors via Imperceptible Background Patches
- Adversarial Reprogramming of Neural Networks
- Consistent Generative Query Networks
- APRICOT: A Dataset of Physical Adversarial Attacks on Object Detection