5 ms·
This is cool work; however, whoever approved this as PUBLICATION should be shot. No software in sight. This is especially bad since "machine learning" is cite
by bsder 10y ago
This is cool work; however, whoever approved this as PUBLICATION should be shot.
No software in sight. This is especially bad since "machine learning" is cited as a cornerstone of the detection. Hardware papers behind paywall or not yet existing (and may never exist). Nothing even remotely useful for somebody to replicate or evaluate.
If this appeared on the web from a company other than Google, everybody would dismiss it as pure marketing fiction.
Maybe I'm being overly harsh, but one of the big problems with 60GHz type stuff is evaluation that what you have is reliably signal. Especially with "machine learning", who knows what the system is latching onto.
- aphextron 10y agoHave you seen the demonstration videos? Some of the applications look really cool and useful, and the hardware is in a pretty advanced state. See https://www.youtube.com/watch?v=H41A_IWZwZI https://www.youtube.com/watch?v=H41A_IWZwZI
- pizza 10y agoI think https://github.com/simonwsw/deep-soli https://github.com/simonwsw/deep-soli is one way to use software with it