Intel Wants To Use AI To Reconnect Damaged Spinal Nerves (engadget.com) 31
Intel and Brown University have started work on a DARPA-backed Intelligent Spine Interface project that would use AI to restore movement and bladder control for those with serious spinal cord injuries. Engadget reports: The two-year effort will have scientists capture motor and sensory signals from the spinal cord, while surgeons will implant electrodes on both ends of an injury to create an "intelligent bypass." From there, neural networks running on Intel tools will (hopefully) learn how to communicate motor commands through the bypass and restore functions lost to severed nerves. The initial interface will use external computing hardware to interpret spine signals. In the long term, the collaborators would like a wholly implanted system to make the connection.
... to its IME. (Score:2)
"Damaged", of course, means "lacking exploitable ... security ...holes", and not the state Intel's backbone is currently in.
Re: (Score:2)
"Damaged", of course, means "lacking exploitable ... security ...holes"
It's going to be "Intel Inside" -- why do you think they want to do it? You just think it's bad when your COMPUTER is hacked, wait until it's YOU.
Ob GitS:SAC ref. (Score:2)
You just think it's bad when your COMPUTER is hacked, wait until it's YOU.
"I thought what I'd do was, I'd pretend I was one of those deaf-mutes"
Learn how? (Score:2)
How will the ANN learn to connect the 2 parts together? What is its training data set? What is the fitness function used to measure success? Its not like it'll be able to tell if the patient is having a piss or walking without outside feedback. Or is this just more from the AI Hype Machine designed to provide a puff peice for - in this case - Intel and contains very little substance.
Re: (Score:3)
How will the ANN learn to connect the 2 parts together? What is its training data set? What is the fitness function used to measure success?
That's why it's a research project, dummy because they are researching how to do it. I mean, if they already had all the answers then it would just be an engineering project.
Re: (Score:1)
Bullwinkle, that trick never works.
I designed electronics for neural implants. Stimulation of nerves is easy. Stimulating them in physical and temporal patterns to provide usable signals is tricky. The signals spread and stimulate more than a few neurons. It also takes a specific amount of charge to trigger the electrochemical senses that will trigger an adjacent nerve signal if the electrodes are so small as to stimulate only a few nerve cells, the current density can be so high that they trigger electroly
Re: Learn how? (Score:2)
I don't think it's possible because it's really difficult.
Seriously ?
Re: (Score:2)
Translation: My first computer had 64KB and now it has 16GB, thus everything is possible. Actual Physics and Science can be ignored because AI.
Re: Learn how? (Score:2)
Posters making statement of opinion-facts should be prepared for some pushback.
Assuming "genius" comments are immune simply reinforces the observations made in the replies.
Re: (Score:2)
Actual Physics and Science can be ignored because AI.
This is literally the motto of "Singularity Club"
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
This is mostly a myth, I worked for a medical imaging company for years. It depends on *where* the implant is. For the most part, ferrous metals just cause imaging artifacts which mess up the image, there's no real danger to the patient.
The only real danger to the patient comes from small, unsecured bits of metal sitting in soft tissue (a bit of shrapnel or steel shaving in your eyeball, a BB in your brain etc..). THOSE can tear loose and cause serious damage.
Most medical implants are secured to bone or fir
Re: (Score:2)
Sounds like its not even a project yet since this is pretty fundamental stuff - more like a load of vague hand waving in order to attract publicity.
Re: (Score:2)
Well if it's like most "AI" and "ML" stuff I've seen being passed off on the market these days,
It'll totally lack the ability to take any user feedback. Which mystifies me as much as it
apparently does you.
Will the tools include Intel's Management Engine? (Score:2)
Everyone wants their spine to be remotely managed, right?
Re: (Score:2)
Well, everyone paralyzed from the neck down probably would be delighted.
That said, this would qualify as a "medical device", which generally require a lot more stability than your average desktop. After all, your desktop screws up, you're annoyed and have to recover from backup, but if your medical device screws up, you're assuming room temperature....
Re: (Score:2)
..., which generally require a lot more stability than your average desktop
And is more difficult. And must be safer from both intentional and non-intentional failures. In other words, if they can't even get their desktop security in order, they should not be allowed to be in the medical business.
Re: (Score:2)
Everyone wants their spine to be remotely managed, right?
I'm putting mine in the Internet of Things or killing myself.
Re: (Score:2)
Everyone wants their spine to be remotely managed, right?
There are already a lot of people like that. They’re called Democrats.
Translation (Score:2)
Intel wants DARPA money and access.
homer simpson (Score:2)
https://www.youtube.com/watch?... [youtube.com]
Done. (Score:2)
Starcraft (Score:4, Interesting)
Do you want Protoss Dragoons? Because this is how you end up with dragoons.
Seriously though, if something like this actually works, you're one more step on the way to where humans could operate machinery as easily as if they were walking. And I don't mean simple exoskeletons but something more on the lines of the mechs from Avatar or The Matrix or the powered armor from Starship Troopers but using the actual nerve signals from the operator as opposed to reacting to their muscle movement.
Re: (Score:2)
Re: (Score:2)
This or something like this has been in a lot of movies.. glad someone's finally trying to make it real (even if we do end up with Dragoons ;)
Re: (Score:2)
...or Doc Oc!
DARPA uses (Score:2)
The obvious use for this that comes to mind is repairing the spine of an injured soldier. But....
Extending the converssation from the other day on designing super soldiers with CRISPR, what if you could install a device at the brain stem that made the spinal cord configurable. A soldier needs to know that he stepped on a nail so that he can treat it, but he may need to continue to function even though he's taken a bullet through a kneecap. What if the soldier could acknowledge the injury, but then turn th
Wait I've seen this one before.... (Score:2)
We actually liked the film but if you're going to go ahead and do RL then no thanks.