Inclusive Speech Technology

This webinar is a conjunction of TU Delft Safety & Security Institute and Campus The Hague.


Inclusive Speech Technology

Automatic speech recognition (ASR) is increasingly used, e.g. in emergency response centers, domestic voice assistants (e.g., Google Home), and search engines. They work (really) well for standard speakers of languages for which there is enough training data to train the systems. However, for around 98% of the languages in the world, there is not enough training data, and consequently no automatic speech recognition systems can be trained for these language. Also, ASR systems tend to not work well for speakers with an accent, young children, people with a speech impediment, and other “non-standard” speakers of the language. In this talk, I will first give a global explanation of what speech is and what information it contains (a lot!) and what you can do with this information (also a lot!) in a global overview of the field of speech technology. Then I will zoom in on automatic speech recognition, and focus on its limitations (e.g., why does your Google Home or Siri not always understand you?) and promises, and explain more about our new research line on inclusive and ethical automatic speech recognition, i.e., the development of ASR systems that work for everyone, irrespective of how someone speaks or the language that person speaks.

Join this webinar


The event is finished.


Jun 02 2021


10:00 am - 11:00 am

Leave A Reply

Your email address will not be published. Required fields are marked *