Speech-to-Text-WaveNet: End-to-end sentence level English speech recognition

This is exactly what I would have wanted for my master thesis about half a year ago, where I wanted to use s2t with good control over the system without having to implement everything myself.

This seems super useful for most speech recognition - understanding context.

It doesn't seem like the mainstream engines (Alexa, Google Voice, Siri) are context aware. Why not?

A few weeks ago, a deep learning researcher at one of the world's leading speech groups told me off-the-record that offline, human-parity speech recognition would be "coming soon" to mobile devices. Not sure s/he realized just how soon that would be. Even though state-of-the-art ASR is really expensive to train, recognition is extremely cheap to run, even on lower-power devices. [1][2] With specialized silicon, you can do this, continuously, for free, on something like a smartwatch. You don't need to open a websocket or call an API running on some beefy server to do this, speech-to-text is now a basic commodity. Fully offline, ubiquitous speech recognition is right around the corner. With human-level speech synthesis [3], speech applications are going to get very interesting, very quickly.

[1] http://niclane.org/pubs/deepx_ipsn.pdf

[2] https://www.ibr.cs.tu-bs.de/Cosdeo2016/talks/invitedTalk.pdf

[3] https://github.com/ibab/tensorflow-wavenet

Also why can't we track emails sent from our iOS device like we can with desktop GMail plugins??

Looking at the training loss graph, it looks like training for more time would produce even better results...

Anyone want to volunteer a few weeks of GPU time to train this better?