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

Neural network acoustic models have significantly advanced state of the artspeech recognition over the past few years. However, they are usuallycomputationally expensive due to the large number of matrix-vectormultiplications and nonlinearity operations. Neural network models also requiresignificant amounts of memory for inference because of the large model size.For these two reasons, it is challenging to deploy neural network based speechrecognizers on resource-constrained platforms such as embedded devices. Thispaper investigates the use of binary weights and activations for computationand memory efficient neural network acoustic models. Compared to real-valuedweight matrices, binary weights require much fewer bits for storage, therebycutting down the memory footprint. Furthermore, with binary weights oractivations, the matrix-vector multiplications are turned into addition andsubtraction operations, which are computationally much faster and more energyefficient for hardware platforms. In this paper, we study the applications ofbinary weights and activations for neural network acoustic modeling, reportingencouraging results on the WSJ and AMI corpora.

Toward Computation and Memory Efficient Neural Network Acoustic Models with Binary Weights and Activations