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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

The Adiabatic Quantum-Flux-Parametron (AQFP) superconducting technology hasbeen recently developed, which achieves the highest energy efficiency amongsuperconducting logic families, potentially huge gain compared withstate-of-the-art CMOS. In 2016, the successful fabrication and testing ofAQFP-based circuits with the scale of 83,000 JJs have demonstrated thescalability and potential of implementing large-scale systems using AQFP. As aresult, it will be promising for AQFP in high-performance computing and deepspace applications, with Deep Neural Network (DNN) inference acceleration as animportant example. Besides ultra-high energy efficiency, AQFP exhibits twounique characteristics: the deep pipelining nature since each AQFP logic gateis connected with an AC clock signal, which increases the difficulty to avoidRAW hazards; the second is the unique opportunity of true random numbergeneration (RNG) using a single AQFP buffer, far more efficient than RNG inCMOS. We point out that these two characteristics make AQFP especiallycompatible with the \emph{stochastic computing} (SC) technique, which uses atime-independent bit sequence for value representation, and is compatible withthe deep pipelining nature. Further, the application of SC has beeninvestigated in DNNs in prior work, and the suitability has been illustrated asSC is more compatible with approximate computations. This work is the first todevelop an SC-based DNN acceleration framework using AQFP technology.

A Stochastic-Computing based Deep Learning Framework using Adiabatic Quantum-Flux-Parametron SuperconductingTechnology