Neuromorphic Computing in Speech Recognition Using Nano-devices

Vishva Bhatt, Hetvi Shah, Kathan Shah, Jigarkumar Shah, Manan Shah

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

Neuromorphic computing has become very essential with increase in human demands for advanced computation systems. The paper intends to take cues from the functioning of neurons as memristors to replace the conventional Von Neumann architecture with data-driven cognitive computing. The main objective is to add the dynamics of magnetization in spin electronics such as spikes. Moreover, the paper aims to integrate and connect the recognition system with nano-scale oscillators to make bio-inspired artificial synapse devices that can recognize speech. This will mean making nano-neurons using nano-devices. A nano-scale oscillator or spintronic is a device that plays a crucial role in the audio recognition process improving it by 70–80%. Furthermore, it reduces error factor and is on a par with complex electronic or optical systems that are used to replicate the cochlear model. This can be observed by control trials where input audio signals, after filtering to frequency channels like the cochlear model or spectrogram, are made to pass through pre-processed input without a spintronic oscillator. The output is directly reconstructed by the computer. The study demonstrates that nano-magnetism and spintronics can be used to perform complex cognitive tasks with key applications in various fields through its speech recognizing capabilities. By using the neuromorphic chips which recognize speech and consume less power, robots will be able to work efficiently and respond to the spoken commands. It will also make smartphones better to understand and anticipate one’s intentions needs.

Original languageEnglish
Title of host publicationAdvances in Systems Engineering - Select Proceedings of NSC 2019
EditorsV. H. Saran, Rakesh Kumar Misra
PublisherSpringer Science and Business Media Deutschland GmbH
Pages45-53
Number of pages9
ISBN (Print)9789811580246
DOIs
StatePublished - 2021
Event43rd National Systems Conference on Innovative and Emerging Trends in Engineering Systems, NSC 2019 - Roorkee, India
Duration: 6 Dec 20198 Dec 2019

Publication series

NameLecture Notes in Mechanical Engineering
ISSN (Print)2195-4356
ISSN (Electronic)2195-4364

Conference

Conference43rd National Systems Conference on Innovative and Emerging Trends in Engineering Systems, NSC 2019
Country/TerritoryIndia
CityRoorkee
Period6/12/198/12/19

Keywords

  • Memristors
  • Neuromorphic computing
  • Speech recognition
  • Spintronics

Fingerprint

Dive into the research topics of 'Neuromorphic Computing in Speech Recognition Using Nano-devices'. Together they form a unique fingerprint.

Cite this