Slashdot Mirror
IBM's Watson AI Implanted Into a Robot, Evolves, Can Now Sense Emotions (hothardware.com)
bigwophh writes that IBM's Watson cognitive computing platform "is now more capable and human-like, especially when encapsulated in a robot body." An article from Hot Hardware reports that this week at NVIDIA's GPU Technology Conference, "We saw Watson in robot form respond to queries just like a human would, using not only speech but movement. When its dancing skills were called into question, the robot responded by showing off its Gangnam Style moves." After winning Jeopardy's million-dollar championship in 2011, Watson moved on to "more practical applications" like providing data-analyzing services for doctors, lawyers, and other professionals, and "the capabilities of what IBM has created are nothing short of amazing... Just like a real person, the underlying AI can get a read on people through movement and cognitive analysis of their speech. It can determine mood, tone, inflection, and so forth."
How long until I have one for carpool lanes?
In a world of the blind, the one-eyed man is king--and the two-eyed man is a heretic.
Let me know when it picks the winning lottery numbers. Then I'll be interested in purchasing.
"the capabilities of what IBM has created are nothing short of amazing... Just like a real person, the underlying AI can get a read on people through movement and cognitive analysis of their speech. It can determine mood, tone, inflection, and so forth."
I'll believe it when I see it.
Thus far, emotional analysis is an over-hyped category and I am getting tired of marketers beating that dead horse over and over again.
The software codes have no ability to sense whether it is a coffee can or a Abrams M1 Tank that hosts the code.
This is just another example of, "My Cat Tricked Me, Therefore It is a Genus!".
Ha ha ja ja
open the damn pod bay doors when I want it to?
When if get's to smart will it try to kill the people who it's feels are trying to trun it off.
but seriously, what does "Gangnam Style moves", have to do with so called AI? such physical dexterity is mainly a feat of mechanical engineering , and can be ( and are) performed by robots without "AI".
Is it likely to explain to us soon how bad Jews are and how good swastikas can be?
Should have "in humans" as part of that, I misinterpreted that headline completely.
'Skynet dances to "Gangam Style" '
That's how they be.
They hate us.
Winning Jeopardy! was indeed an amazing achievement. Certainly more that Google searches, but much less than real intelligence. But that is quite a different type of problem from diagnosing medical issues (Think Mycin, 1980s) or making a robot dance.
So anything vaguely intelligent that IBM does automatically seems to be labelled "Watson". And the ignorant press just naively swallows it.
This is not helpful. There are real AI technologies, and they are different and have different abilities that go beyond a collection of buzz words. Understanding what those really are is interesting. But all we see is either complete woffle in the popular press or very detailed, narrow technical papers from researchers.
http://www.computersthink.com/
is an attempt to address this. Mention this post if you would like a free copy.
They hate science.
Can you fuck it?
Anons need not reply. Questions end with a question mark.
... if Watson learned the Gangnam Style moves by watching videos, and independently decided when using them is appropriate, that would be extremely significant. I am assuming that is not the case. However, with advances taking place in deep learning, AIs may be doing just that in a few years.
It is the fault of others.
If you go to IBM's labs and try to turn it off the police will take you away. If the machine becomes really smart it will convince people that it is important for national security and then be surrounded by armed guards.
But that is all nonsense. Winning Jeopardy! was a milestone, but far from the end of the journey. And the rest is just marketing hype.
Have a look at
http://www.computersthink.com/
For a more considered view. (Mention this post for a free copy.)
More precisely, you could ask "What do Gangnam Style moves have to do with a journalist's for-profit article about AI?", and you might have a clearer answer.
Flippancy aside, they're specifically researching emotional AI, and dancing has long been associated with emotional response in humans. Baby steps, but if machines and people are going to interact more smoothly in the future, it'd help if the software had some understanding of the causes and effects of human emotional reactions.
Why would anyone engrave "Elbereth"?
Always.
Because republicans have no emotions
Which means they don't care about the people. Don't care about the people b
Sounded extremely scripted. Electronic Ventriloquist-Dummy/Yak-bak territory
Here's a parrot that can do the same act except with more cognition under the hood:
https://www.youtube.com/watch?v=bl7WljhLa7Y&nohtml5=False
Has anyone actually played with the IBM Watson APIs? In a word: underwhelming
http://dialog-demo.mybluemix.net/
After the commercials I was expecting Watson to exceed the cleverbot nonsense I played with ages ago. Not even close. From what I can tell, it's just a framework for parsing dialog to nearest neighbors at best, or worse: a lookup table of case insensitive challenge-response pairs. I see zero difference between this demo and a phone tree. At least when I ask a phone tree if it's a computer these days: it says "of course not!"
Anyway, I was so annoyed with the low quality I spent an entire evening looking for the best on the market.
The two that left an impression were as follows:
Best: mitsuku.com
Close second: skynet-ai.com
Mitsuku will take offense to mean things you say to her and take a disliking to you. These are much more impressive than the markov chain based bots all over twitter that produce almost-english gibberish. Supposedly Google's Seq2Seq is pretty bleeding edge. Using movie dialog subtitles as training data is one of the common inputs.
Here's some interesting demos:
http://arxiv.org/pdf/1506.05869v1.pdf
They don't care about home food.
And why does it sound like a woman when it's name is "Watson"
without getting poop on it's fingers? Because that really would be progress.
That's how those pukiabz be.
Socliaisdyt be bout feelz.
Blood and Ed f
I it I'd jaws dey stops uz
Dis
Hodebmysekfbinx cod find
Ratz we wuz feyz b
It must be cold.
They hang Uz upside down.
Bits us needs come outs
Financial analysts overall suggest Watson has been ho-hum as an investment. It may be worth more in PR than in actual products.
Table-ized A.I.
One of these days Watson will be used to monitor election debates..
I do love a good party.
A vampire. Are u kidding me?
By sanders b
Cox day hatz andersonzbbbbb.
Buts Rey's Getz vote 2 aticanz.
Bernie wants us to die v
Only a Virgin born at the minicams of lights be born.
Sailors dean ffonrzbbbb
Let me know when it picks the winning lottery numbers. Then I'll be interested in purchasing.
It did not "Win" the jeopardy game. Others knew the answers; Watson was just set up to buzz in faster than the other contestants.
On vex searched a Finn buttz money.
But notz dere wayz.
Fancy telling u everything.
Oh car lebbbb
Hideki!
It's probably already a lot smarter than you, judging by the number of errors in your one sentence.
I could have sworn Slashdot ran a feminist rant a couple days ago about how AI's are always female 'servants'...
Looks like they forgot to mention two of the most famous ones.
I would program Watson to take everything asked as some kind of sexual reference. ;)
The level of mindless fear, ignorant suggestions and blind stupidity exhibited in the comments here are enough to convince anyone we'd be better off if the machines did take over. You imbeciles deserve it.
You must use this fear. It is your strength. together we will rule the galaxy.
Some drink at the fountain of knowledge. Others just gargle.
Gotta love Slashdot AC telling Watson team they don't have a clue compared to him.
"Artificial intelligence is sort of the holy grail of computing,
and while we may never reproduce the human brain or it's capabilities in their entirety in electronic form"
Stopped reading at this point, as the author has assumed that the "sort of holy grail of computing" is not possible.
The human brain is not magic. I assume the author thinks otherwise.
SCI-FI generally take as a rule that an "AI will never understand human emotions" and thus write em like cold emotionless machines that crush human skulls with their metallic feet, but in reality, it's probably not that hard to manipulate humans by their emotions.
Which means the Skynet probably will take the form of a friendly but horribly manipulative virtual creature that will make people commit suicide by just saying and showing the right things instead of wasting time and resource building robots.
That's just Joe. He's not dumb, I don't think. He does spell very poorly but, until this post, he had his spell-check on as of late. It took some work but he had it on. It was a huge improvement and he must have turned it off - maybe he's using a different browser or computer?
Anyhow, Joe, turn your spell check back on. Watch those red squiggly lines.
Seriously, he says smart stuff on a regular basis. He know a bit about computers. He just doesn't spell worth a fart and his grammar is worse than I've seen from ESL speakers who actually indicated they could really speak English. It's worse than machine translation, honestly.
But, he tries and he's not dumb. Not at all. He just can't write. I believe he can read, he is usually topical. Well, if you can interpret it then he's topical.
Seriously, Joe. Turn your spell check back on. I've been bragging to other people about how well you were doing. *sighs* I'll post this as an AC, I usually remember to. Always the carrot, never the stick. ;-) C'mon Joe, you were doing so good and you've been a bit off for about two days now. C'mon... We're rooting for you. We really are.
Just make sure you bring your space helmet with you always.
Table-ized A.I.
*cough*
I use it every week to cook on weekends.
https://www.ibmchefwatson.com/...
You enter what you have in the fridge and pantry and it suggests meals to cook with fantastic combinations, that real chefs around the world are using too.
Been following this Watson stuff for quite a while, and I still can't say which. Probably not magic, however. Even watched people "playing" with the Japanese Pepper robot in Softbank, and can't figure out if there is anything there. The users may not be sufficiently easily amused?
Freedom = (Meaningful - Coerced) Choice != (Speech | Beer^2), and sad sock puppets' bad mods avail them naught.
What happens when 'artificial intelligence' starts saying what everybody knows to be true, but which the JEW prevents us from talking about? That white people have the right to simply live around their own kind, and the right to NOT associate with somebody beats the desire of somebody else to force themselves into association with YOU...
There must be some limits imposed on how incorrectly words can be used to distort reality.
In other words, PUKE!
Is this the direction we really want to be going in with regard to AI development?ðY- Danger, Danger, Will Robinson!
Not sure if the last line is a joke or not but I would like a free copy of your book, it looks very interesting and if any good, I will highly recommend it to others to buy.
My email is winehacker@gmail.com if you want to send me a eBook.
If they will drive my car for me, I will be the first to welcome our newly intelligent artificial overlords.
No, it will kill the bad spellers first. Low hanging fruit.
When if get's to smart will it try to kill the people who it's feels are trying to trun it off.
No, but it will certainly take out people who use apostrophes like so much table salt.
https://app.box.com/WitthoftResume Code: https://github.com/cellocgw
Did you learn "readings" in school? Did you learn "writings" in school? Did you learn "sciences" in school? Did you learn "histories" in school?
No, you didn't.
You learned reading, writing, science, history and math.
Try to keep up.
AI does not evolve. It executes, you stupid fucks.
It does not think. It does not feel. It simply processes the fucking data.
And it does certainly not sense emotions. It doesn't understand them. It doesn't understand anything.
It just rotates virtual wooden gears in the finite Turing machine to give you somehow valid output.
Fucking learn this already, study ANN and machine learning before you write this bullshit full of magical thinking.
Or just kill yourself. That would be easier and equally beneficial to the world.
OK, this is it. I quit Slashdot. This community is (brain) dead.
Well, it's programmable hardware. In fact, a good bit of it is re-programmable hardware.
And as software can be embodied in either read only or read-write memory, and serves only to configure the hardware as to what steps to take in what order and what priority...
Pretty much the same thing in the final analysis.
--fyngyrz
anon due to mod points
Soon IBM will have achieved it's long sought goal and will be able to fire the remainder of it's non-executive human work force, replacing them with obedient robots. Of course, if Watson is as smart as they say then maybe those robots will be smart enough not to work for zero pay.
That's not actually been established; in fact, the opposite appears to be the case. So far, we've learned that neurons are quite similar to one another; dendrite-to-axon connections similar; brain chemistry essentially uniform; cell metabolism mostly uniform.
It appears to be a challenge of understanding cortical topology more than anything else. Operations such as Numenta's are making inroads in that area.
And again, one doesn't have to completely understand something to use or create it if it has inherent or emergent characteristics of its own. And emergence is something that neural structures bring to the table in spades.
--fyngyrz
anon due to mod points
You seem to think that a real AI project would only use one approach to AI. But of course they don't, any more than a human brain does. Watson undoubtably pulls on data-mining techniques, neural-nets, rule-based AI, scripts and many other techniques.
I am interested in a free copy. How do I "mention this post" to Amazon?
Hah! Can't lift a foot for this demo - look at any human doing this kind of move, they clearly lift their feet alternately off the ground and their system keeps balance, they don't fall over (most of the time, unless they are drunk or stoned).
Now, what about the size of this thing - what kind of message are they trying to send with this? Beats me.
This might finally explain IBM's long game as they purge the company of now unneeded human consultants. It only seemed like a complete hollowing out of all institutional knowledge and ability to create customer satisfaction while pursuing a short sighted focus on quarterly numbers.
This program was made possible by a grant from the Ultra-Humanite, and viewers like you.
Deyy terk err jerbzzzzz b
Classic AI research endgame, is something that works along the Watson lines actually.
Imo what most people think true AI is, is called Artificial Consciousness.
Watson, while impressive, is... not conscious, is it? Eheheh.
Its a step forward for databases and machine learning, not so for AC.
Whether its a step forward for AI, remains to be seen.
Follow the link and look for the email address therin!
Sounds like a great way to spend the rest of the night!
Blackwood/.gmailDOTcom
AT replaces site
Thanks
eh, the email that had somehow been deleted and was just added back ...
Actually, one of the few accessible semi-technical papers on Watson claims that their main contribution to the field was in utilizing several different techniques and then selecting the best result from them. Not so much at the low level like neural-nets vs rule-based, more at the higher level of different engines.
But Wason is what it is, and that aint everything.