Current browse context:
cs.LG
Change to browse by:
References & Citations
Computer Science > Machine Learning
Title: Relevant-features based Auxiliary Cells for Energy Efficient Detection of Natural Errors
(Submitted on 25 Feb 2020 (v1), last revised 26 Feb 2020 (this version, v2))
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.
Submission history
From: Sai Aparna Aketi [view email][v1] Tue, 25 Feb 2020 17:22:10 GMT (690kb,D)
[v2] Wed, 26 Feb 2020 01:30:08 GMT (690kb,D)
Link back to: arXiv, form interface, contact.