The voice interface represents another big shift in user interfaces, and could be as big a shift as the shift from switches to terminals, the switch from terminals to windowed GUIs, or the switch from full computers to small touchscreen devices.

Unfortunately, right now most of these devices are little more than a telephone for talking to a server in the cloud. That is a big problem for privacy, and presents the user with a false choice between access and privacy. Today's computers are capable of handling a verbal interface without requiring a cloud service to handle all of your requests.

Aaron Chantrill is one of the developers on the Naomi Project, a free and open source voice assistant (https://projectnaomi.com). It is a plugin based system based on Jasper (http://jasperproject.github.io/), and depending on how the user chooses to configure it, can run totally offline.

In this presentation, Aaron would like to introduce people to the Naomi Project, talk about the benefits of a verbal interface, talk about the current architectures of virtual assistants, and ways in which that architecture can be altered to provide better privacy, and also do a comparison of offline speech to text and text to speech solutions, including Pocketsphinx, Mozilla Deepspeech, Kaldi, and Julius. Finally, He will invite people to participate in building and improving Naomi. There is a lot to do and much of it is both engaging and cutting-edge.

Aaron recorded a video of his presentation that you can watch: https://youtu.be/8383A_t7iKQ

The slides are available as well: https://github.com/aaronchantrill/NaomiPresentation/raw/master/Presentation.odp

Due to COVID-19, we will meet virtually for a Question & Answer session via Google Hangouts: https://hangouts.google.com/hangouts/_/calendar/ZHJhc2t5QGdtYWlsLmNvbQ.3j6lkajlu4hovgjs0723eta607?hs=121