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

Winograd is generally utilized to optimize convolution performance andcomputational efficiency because of the reduced multiplication operations, butthe reliability issues brought by winograd are usually overlooked. In thiswork, we observe the great potential of winograd convolution in improvingneural network (NN) fault tolerance. Based on the observation, we evaluatewinograd convolution fault tolerance comprehensively from differentgranularities ranging from models, layers, and operation types for the firsttime. Then, we explore the use of inherent fault tolerance of winogradconvolution for cost-effective NN protection against soft errors. Specifically,we mainly investigate how winograd convolution can be effectively incorporatedwith classical fault-tolerant design approaches including triple modularredundancy (TMR), fault-aware retraining, and constrained activation functions.According to our experiments, winograd convolution can reduce thefault-tolerant design overhead by 55.77\% on average without any accuracy losscompared to standard convolution, and further reduce the computing overhead by17.24\% when the inherent fault tolerance of winograd convolution isconsidered. When it is applied on fault-tolerant neural networks enhanced withfault-aware retraining and constrained activation functions, the resultingmodel accuracy generally shows significant improvement in presence of variousfaults.

Exploring Winograd Convolution for Cost-effective Neural Network Fault Tolerance