Slashdot Mirror
NASA Develops Tech To Hear Words Not Yet Spoken
alex_guy_CA writes "Yahoo News has a story about technology that comes close to reading thoughts not yet spoken, by analyzing nerve commands to the throat. 'A person using the subvocal system thinks of phrases and talks to himself so quietly it cannot be heard, but the tongue and vocal cords do receive speech signals from the brain,' said developer Chuck Jorgensen, of NASA's Ames Research Center, Moffett Field, California. Jorgensen's team found that sensors under the chin and one each side of the Adam's apple pick up the brain's commands to the speech organs, allowing the subauditory, or 'silent speech' to be captured. The story indicates the method could be useful on space missions or other difficult working conditions."
Wow! Combine this with a transmitter and receiver, and you get the ability to have sub-vocal backchannel communication with people--I think it was Gregory Benford who wrote a series of books that featured something like this.
Way better than text messaging.
It's Slashdot's evil twin... SlashNOT
...that system Ender used to talk to Jane? That would be sooo cool. (Now, all I need is an omniscient AI with root access on every machine connected to the Internet...)
I mean - there are a lot of things that I *think* about some of my coworkers, especially during meetings, but I always am able to catch myself right before I say anything. You'd *really* have to watch yourself plugged into that thing!
Isn't that lip-reading technology we had on that Jupiter mission three years ago good enough?
....they can hear you scream in space.
Know what I like about atheists? I've yet to meet one that believes God is on their side.
Now I'll have to break out some more aluminium and extend the tinfoil hat into a full-face helmet.
I wonder if this could benefit Stephen Hawking? Good thing he's got friends at NASA. ;)
The title implies that this technology could predict speech before it is said, but the article explains that it can simply read people's conscious thoughts as they are occurring. Those seem to be two completely different things to me.
the ability to "hear" inaudible conversation is valuable in case there is a lot of outside noide (explosion?) or for psychological reasons (a crew member mutters something under his breath)
This could potentially take a lie detector to a new level - people are likely going to think over their possible responses before replying, and this could be used to obtain those thoughts. Scary.
INTERROGATOR: "I'll ask you once more - Did you kill Mrs. Finkelstein?"
PRISONER: "..."
INTERROGATOR: "Aha! And where did you hide her body?"
<<.. wow .. this device feels kinda neat ... ooooh .. tingly on my adenoids there ... i wonder if it's working ... it would be so cool if it did .. i always dreamed of this as a kid .. oh my god being a kid was great .. i wish i could hug my dad again ..>>
No, these damn things simply DON'T WORK!!
Guys, no matter how geeky and nerdy you want to be, NEVER TAKE THIS INVENTION ON A DATE!
but it'll probably be years before we see it commercially.
Imagine using voice commands to control your computer remotely - you're on a croweded bus, using your cell phone to connect to your house computer, telling it subvocally to turn on the airconditioning in time for when you get home, to turn on the coffee maker and download some work from the office and a movie for later. And no one hears anything, and the only thing they can see moving is your throat. What about dictating a letter on your way home, or other documents?
What about secret service agents? Or the military? No more needing to talk into their sleeves or using noisy radio to give away their position. You could have the conversation turn up on a pda screen, or have an artifical voice piped into ear phones. How cool would that be?
I'm sure there's lots more stuff you could use this for that I haven't even thought of yet, but I'm betting it is still years away.
Better start practicing singing a song in your head to block out the thought police. "Mary had a little lamb, little lamb, little lamb, Mary had a little lamb its fleece was white as snow..."
"Good morning, boss (you clueless moron). What (boring and useless) work have we (pitiful understaffed few) got on the agenda for today (and the rest or our meaningless lives)?"
"I'm not feeling well (I need a beer to numb my brain after working for you all day). Can I go home (pub crawling) early?"
Sports cheating: communicate perfectly to coach when you are not supposed to.
Croc Hunter: sneak up on animals in the wild to research, etc, and communicate with team and not startle animals.
Porn: somehow... someway...
Government: give tech 20 more years and when these signals can be picked up remotely, let FBI tap the signals without a court order because, hey, there is a War On Terror(TM) to fight.
Interrorgation: capture truth someone would have wanted/started to say but then held their tounge at the last second.
Slashdot: this tech + reconition to text + scripting = best chance at first post. Just think about BSD dying, and it's dead!
For instance, you're more likely to simply pick a quarter off the floor than to say, "I am going to pick this quarter off the floor." Whereas, you're very likely to think the sentence, "I should buy some wine on my way home from the market" if that's part of your plans.
Seems to me that this technology could, in short order, discern the verbal sentences we fashion for ourselves as part of our daily thinking. But it won't ever pick up on the million thoughts we have each day that aren't based on words.
If this technology gets deployed, society will have to learn in short order that not every thought is legitimate. My verbalizing the thought to myself, "I am Napoleon" does not necessarily mean that I think I am Napoleon.
One last thought. If we get widespread, cheap deployment of this technology, it will have as big an effect on our lives as the World Wide Web.
I'm generally "Interesting," "Insightful," and even "Funny" here. What the hell happens to me at parties?
Most (all?) people actually "speak" when they think in words. This is most observable while reading.
When you think (or read) "banana" your brain creates the same signals (but at lower magnitude) as if you would say it. Your tongue actually moves while your reading. Experiments with mute people have shown that they actually move their hands slightly, as if they were forming the words, they read, in sign-language.
This technology does not read your thoughts, but the signals send to your vocal system. As it catches the signals before they reach the vocal system, it reads "words not yet spoken". If you speak the words or just think them doesnt mattern the system. However, the system doesnt reads your thougths. If you just imagine a banana, but do not think the word "banana", the technology wouldnt catch it (even if improved), as imaging a banana doesnt trigger a signal to the vocal system.
I have discovered a truly remarkable proof for my post which this sig is too small to contain.
At last I will not have to hear four persons shouting in their phones just around my pen in the open space. And there will not be those mad-looking people talking into the empty space on the street. On the other side, someone who talk other the phone a lot may forget actually produce sounds while talking with somebody nearby.
.. not what you think is printed before your eyes get around to reading it.
"A person using the subvocal system thinks of phrases and talks to himself so quietly it cannot be heard, but the tongue and vocal cords do receive speech signals from the brain,"
Notice the phrase "..talks to himself so quietly.."?
This is NOT the same as "thinks to himself"
i.e.you mouth the words but don't blow air through your airway so no noise is made.
it's not friggin' mind reading..unlike most of the level 5 posts seem to think.
As another poster mentioned, in OSC's Ender novels, as well as in David Brin's _Earth_. As a matter of fact, the latter described the device almost precisely as described here. Brin even thought of some important caveats: given how difficult it is for the average human being to keep their thoughts on track for 0.2ms, the thing is almost impossible to use for more than 0.3ms. (The extra 0.1ms is the length of time it takes to think 'FORKING PIECE OF SHI.zza!@EOF' as you reach for the sensors.)
:)
So don't get too excited, all you ADD, quasi-ADD and just plain procrastinatory slashdotters -- whoever ends up using this tech won't be you.
- undoware.ca
Anyone else here feel that getting those words not yet spoken would be an absolute breakthrough in relationships with their girlfriends? Man that would have saved my butt countless times...
To make a pun demonstrates the highest understanding of a language
I've had a device that could detect (and react) to words I haven't even spoken yet, for years now. I call the device "wife".
Underholdning.info
The way I see it, there are three kinds of silent reading.
The first involves real subvocalization, meaning people that read out loud very quietly and move their lips. I don't want to offend anyone, but I find it kind of annoying when people do that. Don't know why. The number of people that do this is in my experience, a minority and occurs most often in people that are weak readers or intentionally reading very carefully. I actually will do it myself when I am following chemistry lab instructions or sometimes when I read in Spanish, a second language.
The next kind of reading is what I think is the most common. There is no subvocalization of any kind, no tongue moving inside the mouth, no vocal chord movement, but words written are mentally translated first to sound-thought and then comprehended. I'm not sure if sound-thought makes sense but it's the best description I can think of. Sound-thought is then what is actually processed and understood by the brain. This I think is most common because I believe that most people's thought process takes the form of an internal monologue. A notable exception is that I know when I do calculus, there is a lot of visualization, algebra, logic and simple math going on in my head and very little verbal thought.
The last kind of reading I think could be considered a form of speed-reading, though I have never taken any speed-reading instruction. I use this whenever I am not reading for pleasure and the passage need not be read super-carefully. I look at the words very quickly and directly process them without first converting to internal sound-thought. If the passage is divided into narrow columns, I can process a whole line at a time and scan straight down the page without any lateral eye-movement.
The level of comprehension and retention when I do this is worse than with methods one or two, but it's extremely fast. The biggest advantage for me anyway that it has over methods one and two is that I retain parts of the text visually. I have a quasi-photographic memory and when I took a course in US History for example, I was able to read names and dates off pages in visual memory because I read the text in that manner.
Now, to get a little more back on topic....
When I was in high school, I competed in quiz bowl trivia competitions like jeopardy, but the questions are only read aloud, and you can buzz in and interrupt the reader at any time before the question is finished and give an answer. Our team and other very good teams learned ways to anticipate the question.
Some teams' familiarity with the subtle structure and format of the questions gave them such a strong intuition that weak teams would swear they were cheating. You got to know individual readers and how much they will continue to read after you buzz in(sometimes finishing the word, sometimes mid-syllable.) Which brings me to my point; if you watch somebody's mouth very carefully when they speak, as they are ending a word and starting to form the next, it's easy to tell what the next letter out of their mouth will be.
A teammate of mine once rang in after a reader had read the letter "J" and then stopped. He could tell the next letter was going to be D, and there had already been several similar questions in earlier rounds asking for different information concerning the same subject that he could tell pretty certainly what the question was going to ask(questions often repeat subject but rarely answers between rounds) So he answered Holden Caulfield.
The other team wasn't pleased at all.
This skill could be practical in interrogations of criminal/terrorist suspects and is what I first thought of when I read the headline of the story. (A better headline would be Nasa develops computer processing of subvocal speech.) A suspect under a lot of stress may come close enough to giving you a name that you can tell from the shape of their mouth that it probably starts with an L rather than a T before refusing to speak any more, and that might be enough.
Could it be used for people that have their vocal cords destroyed ?
I really hate signatures, but go to my website.
Simply take a vocal sample of the person of interest. Now think your words and with a little quick signal processing out comes the voice of the person of interest speaking your words. Fun at parties and for police mounting "sting" operations. Possibly could render recorded conversations inadmissible as evidence.
"Consensus" in science is _always_ a political construct.
As an educator, I see a nightmare if this technology goes mainstream. Kids will send messages through calculators, which is bad enough. How the hell would you stop this?
-Iowa
"He who laughs last, didn't get the joke."-Cap