EfficientWord-Net: An Open Source Hotword Detection Engine Based on Few-Shot Learning

R. Chidhambararajan (), Aman Rangapur (), S. Sibi Chakkaravarthy, Aswani Kumar Cherukuri (), Meenalosini Vimal Cruz () and S. Sudhakar Ilango ()
Additional contact information
R. Chidhambararajan: School of Computer Science and Engineering, Center of Excellence in Artificial Intelligence and Robotics (AIR), VIT-AP University, Andhra Pradesh, India
Aman Rangapur: School of Computer Science and Engineering, Center of Excellence in Artificial Intelligence and Robotics (AIR), VIT-AP University, Andhra Pradesh, India
S. Sibi Chakkaravarthy: School of Computer Science and Engineering, Center of Excellence in Artificial Intelligence and Robotics (AIR), VIT-AP University, Andhra Pradesh, India
Aswani Kumar Cherukuri: ��School of Information Technology and Engineering, VIT University, Tamilnadu, India
Meenalosini Vimal Cruz: �Department of Information Technology, Allen E. Paulson College of Engineering and Computing, Georgia Southern University, Georgia, USA
S. Sudhakar Ilango: ��School of Computer Science and Engineering, VIT-AP University, Andhra Pradesh, India

Journal of Information & Knowledge Management (JIKM), 2022, vol. 21, issue 04, 1-16

Abstract: Voice assistants like Siri, Google Assistant and Alexa are used widely across the globe for home automation. They require the use of unique phrases, also known as hotwords, to wake them up and perform an action like “Hey Alexa!†, “Ok, Google!†, “Hey, Siri!†. These hotword detectors are lightweight real-time engines whose purpose is to detect the hotwords uttered by the user. However, existing engines require thousands of training samples or is closed source seeking a fee. This paper attempts to solve the same, by presenting the design and implementation of a lightweight, easy-to-implement hotword detection engine based on few-shot learning. The engine detects the hotword uttered by the user in real-time with just a few training samples of the hotword. This approach is efficient when compared to existing implementations because the process of adding a new hotword to the existing systems requires enormous amounts of positive and negative training samples, and the model needs to retrain for every hotword, making the existing implementations inefficient in terms of computation and cost. The architecture proposed in this paper has achieved an accuracy of 95.40%.

Keywords: Deep learning; hotword detection; one-shot learning; Siamese neural network (search for similar items in EconPapers)
Date: 2022
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.worldscientific.com/doi/abs/10.1142/S0219649222500599
Access to full text is restricted to subscribers

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:wsi:jikmxx:v:21:y:2022:i:04:n:s0219649222500599

Ordering information: This journal article can be ordered from

DOI: 10.1142/S0219649222500599

Access Statistics for this article

Journal of Information & Knowledge Management (JIKM) is currently edited by Professor Suliman Hawamdeh

More articles in Journal of Information & Knowledge Management (JIKM) from World Scientific Publishing Co. Pte. Ltd.
Bibliographic data for series maintained by Tai Tone Lim ().