Relevant-features based Auxiliary Cells for Energy Efficient Detection of Natural Errors

Abstract

Deep neural networks have demonstrated state-of-the-art performance on many classification tasks. However, they have no inherent capability to recognize when their predictions are wrong. There have been several efforts in the recent past to detect natural errors but the suggested mechanisms pose additional energy requirements. To address this issue, we propose an ensemble of classifiers at hidden layers to enable energy efficient detection of natural errors. In particular, we append Relevant-features based Auxiliary Cells (RACs) which are class specific binary linear classifiers trained on relevant features. The consensus of RACs is used to detect natural errors. Based on combined confidence of RACs, classification can be terminated early, thereby resulting in energy efficient detection. We demonstrate the effectiveness of our technique on various image classification datasets such as CIFAR-10, CIFAR-100 and Tiny-ImageNet.

arXiv comment16 pages, 3 figures, 6 tables

References in corpus (7)

• Ensemble Adversarial Training: Attacks and Defenses
• A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
• LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop
• A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
• Distance-based Confidence Score for Neural Network Classifiers
• Interpreting the Predictions of Complex ML Models by Layer-wise Relevance Propagation
• Natural and Adversarial Error Detection using Invariance to Image Transformations