Comment by thesz
2 months ago
This looks no harder than to train custom Kaldi (circa 2017) phoneme model on brain waves, and using remaining Kaldi's functionality for everything else, except for text-to-speech. There was WaveNet for the TTS at that time, with sound quality that is good enough for (and can be improved by) radio transmission.
Thanks for a link!
point is - such tech is used right now to neutralize individuals. imagine hearing word "bread" inescapably couple hundred times a day coming from an unknown source right to your head. for months and (!) years right at the moment you are trying to conceptualize slightly harder thought than usual. everywhere you go 24/7. while there's no help from anywhere (police hasn't answered me for 2 years and counting) as the general public brushes it off as schizophrenia (it's not - voices completely stopped when the lightning storm took out the electricity) and Church paints it as the second coming of Christ (or antichrist when more suitable).
my mostly uneducated guess of what's going on is: radio wave gets sent, human body slightly modulates it and same signal gets received back and used to reconstruct (approximation of?) EEG from noise delta. neural models is the secret sauce that makes such signal processing possible