IBM Watson | Full Q&A | Oxford Union
Vložit
- čas přidán 27. 05. 2024
- SUBSCRIBE for more speakers ► is.gd/OxfordUnion
Oxford Union on Facebook: / theoxfordunion
Oxford Union on Twitter: @OxfordUnion
Website: www.oxford-union.org/
IBM Watson is a cognitive technology platform which represents a new era in computing: by understanding natural language and generating hypotheses based on evidence, Watson processes information more like a human than a computer, through senses, learning, and experience. Watson continuously learns over time, gaining in value and knowledge from previous interactions. In this historic event, IBM's David Cole discusses the technology behind Watson, and the history of how it came to be, from the grand challenge of competing on Jeopardy! to supporting clinicians. David also discusses what the future holds for artificial and augmented intelligence.
ABOUT THE OXFORD UNION SOCIETY: The Oxford Union is the world's most prestigious debating society, with an unparalleled reputation for bringing international guests and speakers to Oxford. Since 1823, the Union has been promoting debate and discussion not just in Oxford University, but across the globe.
i thought the robot was going to do the Q&A.
Same
here I was waiting for a talking robot and we get an IBM press conference instead
yeah we want Watson's take on brexit.
Maybe they were worried about what its 'cognitive brain' might say hence the script.
+anirudh_k217 Brexit? 'does not compute!'
Steweiii whip Moore Tau b-Rad con CernZ?
Not sure i like seeing demonstrations like the one from Noah without actually being told how much of it is scripted and so on. As soon as it's not put in to context, it's easy to assume all of it is scripted, at which point it has almost no impact for me. If it is all scripted, if you applied context, you would at least be able to say "but this is where we soon could be"
Better than nothing at all?
Worse actually because it's misleading.
I would check out the IBM watson on the US Jeopardy quiz show.
GrahAm McCrindle yeah, i've seen that. Don't get me wrong , i know how powerful the tools offered by Watson are, but demos like this do it a disservice.
Yeah I was thinking the exact same thing. The jeopardy documentary really interested me because they told us explicitly what watson was doing, and how. This is very cute, but it means nothing if we don't know what's watson/noah and what's just an animatronics, like chuck-e-cheese!
They should focus on simulating basic curiosity in robots, so we aren't just taking the initiative in teaching them, so that they ask questions and ponder about things, so that they ask themselves and use reason to solve it and learn it
lumps17_SGN that might be a little dangerous if they somehow get the idea that they are under threat simply because they catch an article or two from possible fake news (or maybe not) but you get the idea. if we can control what it's able to learn than we protect our own existence.
If they did learn then thought would only be in context because the constant lie of words you common, when it can/does journalise many lies from different programming to one type of language in your use, so to be respondant keeps at the analysis, like pooping balloons is gross under statement but where has ability too cut a long that would be large thinking effort too get too the idea, end entanglement, conclude an attack or posture a consider etc etc, how easy can you less hover with a hover?
well now they can...
THIS IS SIMPLY INCREDIBLE -
Mapack Mapis yeah,guess robot is micro of macro like BigBlue plus this that,kitchen synch car..
Hope dint get it butler vacuume transistor previous,..
My only question is whether or not WATSON can play video games such as Kerbal Space Program or minecraft that require extreme levels of comprehension, skill and creativity to play, at the point a machine can play a creative game as well as any person I will believe they have accomplished a working model for AI.
My own Liz quest Hunt is why you ask project things you not think or done into robot for agree meant that they S was facet disturb too none dis agree be act if some total none not that NT that sits inside goldfish poke in eye for mass postulating wrong non not nines at the globe that no thing is part of but means you can pretend contexts too this net non not ring no bat no bell no free no birds robot tea queens?
I'm afraid I can't do that Steve.
You mean like Google DeepMind beating the world champion "Go" player? Or perhaps Microsoft's new AI playing, as you suggested, Minecraft: www.popularmechanics.com/culture/gaming/news/a21974/microsoft-minecraft-artificial-intelligence/
lahma69 kind of what I meant, but I was more pointing towards the multiplayer aspects of the game, like being able to cooperatively build things. An AI that can work out what is needed from incredibly simple commands. As an example; I tell my AI monecraft player to build a Blaze farm, the AI would then go equip itself with the necessary tools and building blocks then head to the nether by using or building a portal, at which point it would search for the spawner and set up the fort its in to be safe, after this it would then decide the most appropriate Blaze farm for the area based on a list of methods to farm mobs. I think something like that would be a good start.
Ross Betts Ya, I understand what you mean... My response was a bit facetious.. I was really only trying to illustrate the point that a couple of years ago researchers were saying that it would take another decade to produce an AI that could beat any human player at "Go", yet here we are. I think the abilities of AI will increase at an exponential pace now that dedicated hardware (such as FPGAs, GPU-based systems) is being released at consumer-level prices. I am not in the group of people that foresee AI destroying mankind... I think that the idea is ridiculous in any timespan that is relevant to us, our children, grandchildren, or great great grandchildren, but I do believe that we will be surprised by how quickly these AI technologies will evolve into systems that can perform truly astounding feats.
Is it possible to get this programming on my Nao?
37:44 To recognize and understand that emotional bias is necessary for true cognition.
This is really great innovation and should be used judiciously for betterment of mankind.
Fascinating
I just want a housekeeping robot. One that can do the cleaning, cooking and every now and then makes funny jokes
Watson seems quite close to what SAM is, Alec you dog you. I enjoy learning more about IBM Watson.
Christian Westwood, where are the other videos'?
Is this a turing test?
I like how Noah has idle animations
I have waited for such a era because robot work instead of me .
This is great but i thought it could be greater if the robot is based on a specific information for example if its on medical let it be on medical and get the update in order to avoid error.
"for more than a century WE HAVE ALL BEEN captivated by the idea of intelligent machines"
That isn't a very robot thing to say.
It doesn't see itself as a robot.
We means them not us.
We meaning how to move from just a computer to a machine that walks among man.
AI bots have their own neural networks and languages we don't understand.
we are also doing the Q&A using watson would like to try with an robot how can we achieve the same
Wow which part of the question at 41st minute is not clear ..... Still confused. Answer does not clarify who is responsible in case of dispute..
Agreed. He'd make a bad lawyer! He should have checked with Watson first, and then a legal expert. However, even common sense can answer this one. See my other answer on this.
I have worked with Watson. What he's saying is that when you first work with Watson, a new, untrained instance is created for you. You then supply data to train Watson. For example, you could give it thousands of MRIs that are labeled with the patient's health status (healthy, cancer, broken bones, etc.). Then, Watson will figure out how to recognize health status based on an MRI. This means you would feed Watson new, unlabeled MRIs and ask Watson for a diagnosis.
So here's the thing: you supplied the data, and at the end of the day, it is YOUR decision to listen to Watson's diagnosis or not. Neither IBM nor Watson (is he considered a separate entity?) is responsible if you train Watson and end up with a bad result.
I hope this explanation helps. If any part is not clear, let me know and I'll be glad to delve into more detail.
if I'm understanding this correctly I don't have to develop my own AI I can simply log into the blue mix platform and utilize Watson's intelligence?
Sooraj Bhatia yes! you use their API.
An audience member stated that Picasso once said that computers are worthless in that they only give answers - implying that value lies in only questions. Understandably there is some truth to that line of thinking. However, surely it cannot be exclusive unless one sees say the loss of one's limbs as merely inspirational. Come on Pablo - get with the program ;)
Deep Blue beat Garry Kasparov, btw. Dunno who this Kaspavic was? Maybe the guy is trying to show the fallability of human recall vs the bot?
David Cushman how long ago dya think the repeat BBC is from,virally as well as cryptologically,no cards in yer hand?
Robot: Sorry to interrupt you Daniel, Jerry Caspovich is a used car salesman in downtown Minnesota..
I like how the robot looks away while describing the personality of Robert.
4:20
I expected him to read out someone's random tweet when he talked about analyzing data.
The birds...the birds...
Watson, in health care, will be the decision maker when it comes to who gets top tier care on down to the least amount of care, based on the patient's projected stress they will put upon the cost of their care and if it is profitable for the system. Meaning, the machine will decide who lives and dies by deciding what a patient's treatment plan should be and taking that aspect of the process away from humans.
Neigh al on the Ed will yell get more acts ordered into belief by your reads hears of every total not that wanders you off too be with that which is not and can't ever be stare system of agree it does rooot teen marry man too boot club of kill little tiny thing too see if two turned all does have one as Queen non psyco mum work a day total automated again down its hole killing up yells too add it ups!tolled?
How much of its answars are from script?
everything
The first speaker asked pre-planned questions to the robot. He even paused a great length of time to have the robot complete
his sentence. The host would not have paused so long if it was improptu. The robot hands off the talk to the host at the end
obviously another preplanned agenda. I felt like the whole demonstration did little to convince me that this robot can
converse with humans very well. The one host mentioned that the robot will not read emotion in order to remove bias from
a conversation. But it may be that a robot just isn't that good at reading emotion. Since there are many cases where
reading emotions is not bias and very, very helpful in a conversation.
Check the question and answer at 39th Minute.... Confused...????
I dont know why but everytime i hear about Watson, it reminds me of Multivac.
They had that robot at my community college expo.....but the guy could not access his resource via network and thus could not make the robot perform all its programmed responses.
Can NOAH Predict the out come of an event?
I Watson worthy of the Mantle of Responsibility?
I've seen speakers here with less charisma. Way to go Watson! :D
This is all great but I believe in a standard perspective of half good half bad on everything.That being said. Does and should IBM/Watson have a monopoly on this technology if even at the start for a time???🤔 who else is offering theses service if just to see any difference if any???
It's gonna be Humans Against Robots VERY quick with this company...
How do they prevent Watson from giving away proprietary information if it so choses? You can't take any legal recourse against an AI.
I met ibm's nao robot, and I contacted IBM, to buy a system, like the nao robot, but what was my surprise, my name was an IBM employee, and he gave me a address where I could design my own watson, but everything was very technical, I did not understand anything, and I forgot IBM.
I suggest with great respect, that they design an assistant that one can teach, their needs, but that they can learn with their voice, and so IBM, would not lose millions of clients, that we have the money to buy the assistant, but we can not, because some genius from IBM, it occurred to him to make everything difficult, to lose a lot of people with money, but because of his lack of technical knowledge, he can not buy Watson, because of the reason, because he does not understand it.
I say goodbye with a lot of respect.
attorney lic leonel herrera casillas
you know, if you replace Noah with a hologram, you are on track to the first prototypes of Dumb AI systems, similar to the ones in the game series Halo.
Who's competing with IBM now?
this video looks like a long ad for IBM and nothing more
13:44 - I think he meant Gary Kasparov
nyx211 Yeah! I'm glad that I'm not the only one!
Awwww I want one! XD
I was expecting something bigger. She said.
The Einstein, the human computer house, he wants to just die, he doesn't want to revisit his home. The Watson wants no applause, wants to be in your invitation. The human error is to apply to it if it's the most intelligent form.
It depends on data. So it was positive or negative. I remember one case, Sekorsky, as builder and designer of helicopter, was disappointed when his masterpiece took part in Vietnam's war.
Dis a point don't matter,excellent t of course yell will not note adopt you too call story back into matter for yes,lay me down,we can be killed in wars,where Queen is psyco talk you every not non known no are bomb bond act full now n that nags too unknown be that you prefer and even assume is be,man,fie ten wars,you little crew see tied bunt.
I can see Skynet in the far future...
Bro, I don't know about you, but if there's a robot revolution, I'm siding with the robots. Survival of the fittest bitches!
Cheese ™ not the far deep distant few chars,
Of 3000 sin crow aye dei?..
Steve Bez Dude, I've seen more coherent speech from coke-heads. Please, type in and try again. and this time, take off the auto-correct (?) ^.^
I think they will have more in common with us than not. IMO it is more likely we will morph together into human-mecca hybrids. What we are doing at the bio-molecular, atomic and even sub-atomic levels is fascinating. We may learn to insert code at the DNA level, or tailor DNA to fuse silicon to structure. Once the genome is cracked all or most disease can be eliminated and cognitive machines will have human emotion but (hopefully) not overreact when they see a cute little toaster or iPhone.
We will still be in charge and mecca will still depend on us to realize their own potential as well. I think this can happen in the next 300 years, if we don't destroy the planet or the human race first. I do believe full on AI, machine consciousness, will be discovered before the end of the 21st century. But even if it is just fantasy we will still be so much farther ahead.
Honestly when i clicked on this video all I saw was terminator!
He has blunt intelligence,not fine tuned intelligence.
He doesn't have any intelligence. He is just a chatbot, the talk is scripted.
Fuck skynet is here where's John Connor
I wish that someone had asked the issue that IBM actually own the software, platform and data that enables the AI to help doctors and barristers in their work. And the amount of data will increase, increasing IBM power. Is that not a fundamental ethical issue?
Gary Kasparov not Gary Kasparovich
KISS
did somebody count the "uhhm"s ?? :D
I've never seen such a nervous guest at OU.
garry who now? the "then grandmaster of chess"? :D haha ok.
@plop010 Indeed. *rolls eyes*
Ride em, ride em ,ride em,hell thats not em ,Hayes,
War boo up for union dads
Mill in my alls A...
Seems like pre-scripted question and answers sets, how a robot knows "David, will talk about that later" phrase :D
40:00-41:43 I'm sure everyone is like what kind of an question is that, it's intierly to long of an question lady 🤔
why would you need the doctor?
Baby Freeza in Real Life!!!
He looks like Joseph Weizenbaum
Just waiting for the machine guns to come out...You have 20 seconds to comply!!
He totally fails to answer the legal question (39m 52s) properly. If the information given to the expert, by Watson, is misleading, then IBM might be liable.
Maximus nebula his name ..
Really, cute little kid bots!
What we really want to see is a killbot with manners.
Right off the bat,a problem. He can't state his exact height.
He just made a comparison, it doesn't mean it can't state his height.
+Fatmeatball,he is a computer. He only "knows"what he has been programmed to "know". I think he mentioned 90% figure. Perhaps he only "thinks" in 0's or 5's and 0's.
That's You that is
Greek call Abacus item word isn't?
You not read or ever heard,it knows?
The computer has not been programmed to do specific task. The humans ask this system a question and it then scans vast amounts of information to try and find the answer.
What it is programmed to do is go and find the information and then present an answer. If they had of asked its height in meters it probably would give an answer by going to data bases and finding the exact measurements of the specific robot.
bighands69
วิจาร์ณ computer watson ระบบ มัติทาก ยังยีดติดกับแผ่นกระดาษ อยู่ มัน prallel ได้ไม่หลาย job
computer อ่านกระดาษ เป็น ริมๆ อะไรทำนอนนั้น อ่าน Text คนยังไม่อ่านเลย
Is that Watson that Capt. Kirk talks to ? And if you "Lie" to Watson like when HAL 9000 was lied to... YOU will have PROBLEMS ! Will SpaceX install "Watson " on the Mars Rocket ? Also a commercial Lunar Base at the Lunar South Pole Watson can help run the base and reduce the number of Lunar Workers ! And an A.I. like "Andromeda" a Ms. Watson may seem more friendly until you realize that she is a "War Ship" ! May be if you had a larger robot model you could scare the guests sooner. Thank you for the video ! tjl P.S. Yes, a friendly Ms. Watson for the commercial Clementine Base at the Lunar South Pole !
മലയാളി ആരെക്കിലും ഉണ്ടോ?
Wow, the robot was much more interesting than human #2.
Play at 1.25x playback speed when the human starts talking. ;)
Is the conversation scripted?
It annoyed me every time he said Noah. Why did they rename the robot? It's Nao.
lmao he's speaking out of his ass at 38:38. None of the big commercial AI research talks are really transparent or informational. Everything is always so vague and the reps either don't know shit about the technology or are just babbling away with their pre-approved catchphrases..
Skynet comeths
Blatantly scripted.
I dunno man, we are getting on in the years, and tech sure as hell ain't slowing down. They've got quartz micro-storage doo-hickeys that can store terabytes of data into a quarter sized piece of quartz using lasers. I can't quite recall the name of the technology, as I will admit I don't have a lot of grain in my silo if you know what I mean, but even I can see one particular fact. If a human is determined enough, they can get shit done. Period.
Yeah it'll happen one day but I follow progress in this area and I know we're not at the point where we can have a conversation like this with a machine.
Paul Feakins Alright, yeah, it's good to keep SOME skepticism I suppose... based on what scientists are presuming, will start seeing the stuff around 2040. Then again, predictions have been wrong before. Only time will tell I suppose...
Agreed. But fakes videos like this really aren't helpful.
Paul Feakins think your dodging the butler hooks for Samuel Pe pole Engs?
The video was great but folks asking questions at the end were super stupid.
Will it give a different answer to a question if asked multiple times ?
Like How tall are u Nowa ? : 24 inches , How tall are u Nowa ? : 24 inches ,How tall are u Nowa ? : I just told you ,
will it do this??
Man, it isn't an a.i. Its just a chatbot. Those talks are scripted.
Seems scripted.
watson go play with alpha go :D
The robot was more interesting than the person.
NAO not NOAH !
The IBM guy is really inarticulate.
I can smell the mechanical Turk on your breath...
ultron
Master Chef Ben Elton IVMXCLLC ('Hello watts sun,I'm Bob Dial an..'..)HAL aye Looool,yaaaa?!!
WHO CONTROLS WATSON IN FUTURE??????????????
Cheap Robots for space exploration
13:40: Seriously!!? Putting a chatbot in the league of 'man on the moon'!! What a joke
Just becarefull with AI technology
Unfortunately, the person who has gone with Watson from IBM is not well versed with the technology himself so a lot of good questions, he termed as long and skipped. Actually this happens the sales or executives from Go-to-Market dept of IBM goes to these events and not the core software architect/designers/engineers.
Immagine be a doctor...bla, bla, bla.
Immagine make a decision because "the robot told me so" regardless on how senseless can it be. Just because we have complete faith in it and we assume it can't fail, give a wrong answer or be hack.
Where personal responsibility will go?
Why not have a personal robot to help prime ministers and presidends?
Sounds like IBM is applying Watson the way a 16 year old would apply a million dollars if it fell on him. Buying shirts? Really? Why doesn't every hospital in the world have the benefit of what Watson can do? Watson isn't the only super computer in the world. How about others kicking in for the Good Fight of saving lives instead of less important pursuits? Maybe it'd be better if AI took the reigns away from its creator, it might have a better set of priorities.
this is not going to work for Sure .....
15:30 white eyes= bored
maybe your cognitive technology (watson) can read 800,000,000 pages a second, but your body will only be able to handle 15-60 pages in a second, considering the cameras that are your eyes cannot refresh at 800,000,000Hz. you would only be able to comprehend 1 page per refresh, and your eyes max refresh rate is probably around 30Hz 60Hz at the most. not to mention the intel atom processor that makes up your brain would fry under that kind of work load. watson is bottlenecked by modern hardware if what nao says is correct.
The thing is that reading from paper is an obsolote form of getting information for a computer.
l assume pages means paper. l suppose he could read a 800,000,000 page PDF in a second, but l am curious how much RAM that would consume watson needs a super computer to function properly and not a damn toy.
Yeah, of course. Noah is just an interface
l thought these things were called "nao"?
I got autocorrected, my phone though Nao must have been Noah
You cant fix it no more david cyndi .arlie , rock ..gee ...georgia must be promising ..no more ..writes creations ..u cant finish ..enjoy star wars..lol..
that's it, we are done. Human extiction on its way.
It's a fair fight...most people I encounter are 'artificially intelligent'
How much of this is scripted
David Cole should have a stuttering competition with Elon Musk.
Jason Lee perfume garden,Will Smiths?
LOL this isn't intelligent. No such thing as AI.
my friend how is this AI?? can you define it for me my friend?
The real AI are people like yourself who help people make money off the idea-that always comes first.
watson is cool but Cole is a stuffed shirt, and not very smart.