Using a word list to help train automated speech recognition engines

In our meeting this week we had a question about captioning services which provide a means of enhancing the accuracy of automated speech recognition engines by using a prepared word list to augment the standard lexicon used by the service. I mentioned seeing this before and wanted to provide an example. Here is some documentation on how Verbit does this:
Best Practice for Customers When Adding Materials in the Platform – Verbit Support Center<>

Specifically, they state:

"Uploading preparation material helps with the output

"Verbit clients have an important role to play in the personalization process.

By sharing idiomatic and distinctive terminology pertaining to the transcription, meeting, or event; Verbit users can train the Automatic Speech Recognition to detect and understand topics and niche terms which in return, deliver users with superior accuracy and increases the chance to receive a customized meeting or event experience.

"To accurately transcribe singular terms, Verbit’s Automatic Speech Recognition requires users to upload a list of people's names, brands, places as well as specific professional terminology ahead of time."

I believe that many other service providers do the same. For example, here is some information from 3Play Media: How to Use Wordlists for Live Auto Captioning Quality | 3Play Media<>

Hope that helps.


Steve Noble
Principal Researcher, Accessibility
Psychometrics & Testing Services


502 969 3088<>


[NSF's Convergence Accelerator - 2022 Cohort Member]<>

Received on Thursday, 22 June 2023 20:48:27 UTC