How to "Talk" to Your Software: Alexa, Google, Watson, and Cortana, a Side-by-Side Comparison of Cloud Speech Recognition APIs
May 2018 • Presentation
With the introduction of products like Siri, Cortana, Alexa, and Echo, speech recognition is now part of daily life. This presentation shows how to pick the right service.
Software Engineering Institute
This presentation was created for a conference series or symposium and does not necessarily reflect the positions and views of the Software Engineering Institute.
The APIs behind the voices of Siri, Cortana, Alexa, and Echo are now available for all to us to explore on the cloud. How to choose which one to use (and when) can be mind boggling in the context of different pricing models, API rating schemes, model accuracy, security, and privacy concerns.
This talk is based on experience working on a startup venture that was betting on natural language processing as a key enabling technology. It is a fascinating journey of exploring the API options available and comparing what you can get with each of them. Although there is some overlap, the research also demonstrates significant differences in approach, accuracy, and use case application. Node-red, an open source flow design tool, is used to illustrate the patterns and tradeoffs of the APIs within the context of a simple use case.