English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep Neural Networks (DNNs) are widely being adopted for safety-criticalapplications, e.g., healthcare and autonomous driving. Inherently, they areconsidered to be highly error-tolerant. However, recent studies have shown thathardware faults that impact the parameters of a DNN (e.g., weights) can havedrastic impacts on its classification accuracy. In this paper, we perform acomprehensive error resilience analysis of DNNs subjected to hardware faults(e.g., permanent faults) in the weight memory. The outcome of this analysis isleveraged to propose a novel error mitigation technique which squashes thehigh-intensity faulty activation values to alleviate their impact. We achievethis by replacing the unbounded activation functions with their clippedversions. We also present a method to systematically define the clipping valuesof the activation functions that result in increased resilience of the networksagainst faults. We evaluate our technique on the AlexNet and the VGG-16 DNNstrained for the CIFAR-10 dataset. The experimental results show that ourmitigation technique significantly improves the resilience of the DNNs tofaults. For example, the proposed technique offers on average 68.92%improvement in the classification accuracy of resilience-optimized VGG-16 modelat 1e-5 fault rate, when compared to the base network without any faultmitigation.

FT-ClipAct: Resilience Analysis of Deep Neural Networks and Improving their Fault Tolerance using Clipped Activation