AI Allows Paralyzed Person To 'Handwrite' With His Mind (sciencemag.org) 9
sciencehabit shares a report from Science Magazine: By harnessing the power of imagination, researchers have nearly doubled the speed at which completely paralyzed patients may be able to communicate with the outside world. In the new experiments, a volunteer paralyzed from the neck down instead imagined moving his arm to write each letter of the alphabet. That brain activity helped train a computer model known as a neural network to interpret the commands, tracing the intended trajectory of his imagined pen tip to create letters. Eventually, the computer could read out the volunteer's imagined sentences with roughly 95% accuracy at a speed of about 66 characters per minute -- twice the previous record.
You Train Both in Conceptuilisation (Score:3)
To do it much faster, you need to train both the individual and the thought pattern reader, in the concept of conceptualising the whole message and then deciphering that thought. So hunger becomes I am hungry and then more concept thought state models are created by the person and interpreted by the machine, out into long slow speech. Both the person and machine need continuous feedback training, so I am hungry and feel like a banana, out of banana hunger thought and expanding from that ie type of banana how ripe how large.
Use stenography (Score:1)
FIRST (Score:2)
Re: (Score:1)
Why not just type? (Score:2)
If you are looking for hand motion patterns related to writing why not just go with hand motion patterns related to typing which is far more common and high speed these days. Seems like the motions are highly repeatable, faster, and far more consistent than the motions related to individualized handwriting. Not selecting letters off a screen or moving a cursor to letters, but actually typing on a keyboard in your mind.