Estimating the Brittleness of AI: Safety Integrity Levels and the Need for Testing Out-Of-Distribution Performance
arXiv:2009.00802
Abstract
Test, Evaluation, Verification, and Validation (TEVV) for Artificial Intelligence (AI) is a challenge that threatens to limit the economic and societal rewards that AI researchers have devoted themselves to producing. A central task of TEVV for AI is estimating brittleness, where brittleness implies that the system functions well within some bounds and poorly outside of those bounds. This paper argues that neither of those criteria are certain of Deep Neural Networks. First, highly touted AI successes (eg. image classification and speech recognition) are orders of magnitude more failure-prone than are typically certified in critical systems even within design bounds (perfectly in-distribution sampling). Second, performance falls off only gradually as inputs become further Out-Of-Distribution (OOD). Enhanced emphasis is needed on designing systems that are resilient despite failure-prone AI components as well as on evaluating and improving OOD performance in order to get AI to where it can clear the challenging hurdles of TEVV and certification.
References in corpus (16)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Deep Speech 2: End-to-End Speech Recognition in English and Mandarin
- Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
- Domain Adaptation for Statistical Classifiers
- Adversarial Attacks and Defences: A Survey
- Generalisation in humans and deep neural networks
- Xception: Deep Learning with Depthwise Separable Convolutions
- Self-training with Noisy Student improves ImageNet classification
- Does Distributionally Robust Supervised Learning Give Robust Classifiers?
- A Fourier Perspective on Model Robustness in Computer Vision
- Fixing the train-test resolution discrepancy: FixEfficientNet
- Pretrained Transformers Improve Out-of-Distribution Robustness
- Benchmarking Adversarial Robustness
- ComDefend: An Efficient Image Compression Model to Defend Adversarial Examples
- Anomalous Example Detection in Deep Learning: A Survey
- Adversarial Examples for Cost-Sensitive Classifiers