Lessons from Our Voice Engine: Keyword Detection
June 3, 2021
VP of Speech Technology, Dr. Amelia Kelly, brings you the sixth installment of our “Lessons from Our Voice Engine” series, featuring high-level insights from our Engineering and Speech Tech teams on how our voice engine works.
What is keyword detection?
Keyword detection is a feature of the SoapBox engine that allows users to search for a word or phrase (even multiple examples simultaneously) within an audio file.
Why is it important for our voice engine and kids’ speech recognition?
In kids’ speech applications, we may only be interested in a particular word or phrase that is spoken by the child. For example, this word or phrase may be the response to a multiple choice question or a password that triggers an action within the platform, game, or application.
Keyword and keyphrase detection allow the client to focus only on the voice information they’re interested in, without requiring them to overstructure the question or limit the child to a one-word answer.
What else should people know about keyword detection?
Keyword detection is a great feature for any voice-enabled product or service as it encourages kids to interact freely using their voice, while not overburdening the client with too many unnecessary data points. It also delivers incredibly high accuracy for the keywords and phrases in question and works off-the-shelf.
Continue to Lesson #7 on intent recognition, or catch up on our previous “Lessons from Our Voice Engine”:
- #1: Natural Language Processing
- #2: Custom Language Models
- #3: Deep Learning
- #4: Debiasing
- #5: Acoustic Models