Honkling: In-Browser Personalization for Ubiquitous Keyword Spotting

Abstract

Used for simple commands recognition on devices from smart speakers to mobile phones, keyword spotting systems are everywhere. Ubiquitous as well are web applications, which have grown in popularity and complexity over the last decade. However, despite their obvious advantages in natural language interaction, voice-enabled web applications are still few and far between. We attempt to bridge this gap with Honkling, a novel, JavaScript-based keyword spotting system. Purely client-side and cross-device compatible, Honkling can be deployed directly on user devices. Our inbrowser implementation enables seamless personalization, which can greatly improve model quality; in the presence of underrepresented, non-American user accents, we can achieve up to an absolute 10% increase in accuracy in the personalized model with only a few examples.

Howl: A Deployed, Open-Source Wake Word Detection System

“…This training process should take less than a few hours on a GPU-capable device for most use cases, including ours. Finally, users may run the model in the included command line interface demo or deploy it to the browser using Honkling, our inbrowser keyword spotting (KWS) system, if the model is supported (Lee et al, 2019).…”

“…However, our previous line of work demonstrates the feasibility of in-browser wake word detection with Honkling (Lee et al, 2019). Our application is written purely in JavaScript and supports different models using TensorFlow.js.…”

“…To this end, we have previously developed Honkling, a JavaScript-based keyword spotting system (Lee et al, 2019). Leveraging one of the lightest models available for the task from Tang and Lin (2018), Honkling efficiently detects the target commands with high precision.…”

BrowVis: Visualizing Large Graphs in the Browser

“…As a consequence, a recent stream of work focuses on building high-performance data analysis and management systems that run completely in the browser. A limited list of examples include the following: El Gebaly and Lin [14] present an analytical relational DBMS implemented in JavaScript that runs within the browser; Lin [15] describes a self-contained JavaScript-based search engine; Lee et al [16] propose a JavaScript implementation of a keyword spotting system that can be deployed directly on user devices.…”