IBM Watson: Final Jeopardy! and the Future of Watson
ฝัง
- เผยแพร่เมื่อ 5 ต.ค. 2024
- After competing against the two greatest Jeopardy! champions of all time, the technology behind Watson will now be applied to some of the world's most enticing challenges. Watch a breakdown of the match from Ken Jennings, Brad Rutter and the IBM team members as they look toward the future.
Visit ibmwatson.com for more information.
I think you're missing the reason for Watson being built...
What's important is NOT how many people it took to win, but the fact that the people were able to build a MACHINE that's capable of winning
In the last days knowledge will increase, without love, all said and done is as nothing as first Corinthians chapter thirteen says in holy Scripture.
@@davidnewhouse5447 lmao your silly god delusion didn't make computers think, engineers did.
@@davidnewhouse5447we’ve been in the “last days” for over 2000 years, nobody believes you guys anymore
It is not a win for machine. But it is a win for the minds of the people who has developed algorithms to make m/c mimic human thinking
Hahjsjs
Let your customers get satisfaction.
*****
AI is build by human. Algorithms are developed by human. So Welcome to AI...M/c is not human.
+Ram Nangunoori the firs human to be "purified" for his blasphemy in 10 to 15 years. :D
@@FollowersOfPrabhupada Human programmers don't give it the knowledge, the computer has a code that allows it to learn.
Watson: IBM can I have neural network for medicine?
IBM: For medicine?
Watson: Yes.
Actually plays Jeopardy instead
_Money time_
@creditsunkown7974
Unfortunately, no. Waston X is now working in medicine and logstics instead if acquiring the bag 😔
Watson had 36 different operating modes, to switch between calculating a question, waiting, responding, etc. If it switched to a wrong mode or didn't switch properly, it could've not even answered and blanked out.
i for one welcome our new computer overlords
+Roy Staggers Have you ever heard* a computer insult people for no reason?
Roy Staggers vwnsid
Roy Staggers vwnsid
Irving Bisman
Pink
it's called natural language processing. It's not a simple matter of entering a shitload of encyclopedic facts into a computer. But being able to understand the question in Jeopardy's format. Trust me - in computer science this is huge.
@geosunkist
True
I am an undergraduated computer engineer, and i have to say, IBM researchers are amazing, i REALLY would like to have teachers with half skill of this team, thank you for making such a great thing, is really amazing how WATSON can handle data, improving the machine capabilities is a half way to build a smarter planet
well done ibm and the watson team. this is truly an amazing achievement, I can't wait to ask watson a question via the web ;-)
@thesystemsucks
"Skibidi"
I'd be interested to see how it got a 10% confidence in "Omaha" as an answer.
I know this is older, but it's so good. The IBM team talks about their computer contestant on Jeopardy.
So how much longer until the Hal 9000 is released?
3 years
+Marcus thompsct eon
frogs hth hth rg disgusted data A ooh time uts rg Check nfdsdddsddddddddddddfffshtdd fg h grjj dfrydyyyyhhhhgtfatteueieuedh eh g wet t etc tee wet rg h iffy err r eh err dry cm ggh dsdeeertyisdjks bn is gbv
Unun Pentium chgo8005
evilmick66 p
This is just fantastic! Well done to the team.
Jeopardy is nothing more than facts memorization, like trivial pursuit. It is hardly a true measure of all parts of intelligence, in fact it would measure only the least important parts. It is more closely related to a calculate than anything else. The worrisome thing about Watson's failure in Final Jeopardy is that it gave a factual answer that even this simplistic grid should have ruled out. Specifically, it is simply a FACT that Toronto is not located in the US and so any intelligent 5th grader would have ruled that possibility out, and left it blank or guessed an American airport at random rather than put down an answer that was ruled out. Thus, the breakthrough for Watson was not in the "intelligence" portion of its machine, but in its processing data (in its "hearing" it, etc.). But its flub in final showed more how FAR IBM has to go rather than anything else.
This really is amazing. As the guy in the video said, this really shows us how complicated the human brain is. A huge team of researchers, many hours/days/weeks/months/YEARS of work, and it is still not quite on par. Yes, it won Jeopardy, but I'm wondering how it would fair in other things.
An historic moment. Watson is our future.
Well done to everyone at IBM: this is a milestone achievement.
"Watson.. the board is yours..."
"SKYNET.. ACTIVATED..."
great -- we're supposed to root for the computer to beat the humans? "Boo, John Henry!" Yay, Terminator!
It took an army of computer engineers to build a robot that won in jeopardy against 2 humans that held their own. Great accomplishment, no doubt. However, how greater an accomplishment is the human brain didnt have an army of engineers building it? This whole experience is a 2-fold wonder! :)
It's the beginning of skynet.
Swisgard Toki The beginning of skynet was the founding of google.
no team of engineers... just One.
Viktor Cheban A team of engineers over a few years. Humans: evolution (what amounts to a toddler hitting random buttons) over hundreds of thousands of years. At Watson's rate of growth humans will be blown out of the water.
Nubro Zaref . el
@thewinrar2 Toronto's largest airport is named after a Canadian Prime Minister who served during WWI and WWII, and the second largest airport is named after a WWI hero (Billy Bishop). It's not a terrible guess, considering how confused Watson was by the question (ignoring the fact that it completely skipped past the category name)
I'm proud to work at IBM -- what a great achievement as we celebrate our 100th anniversary.
Fast forward to 2022. What is IBM doing now? We have enormous breakthroughs in performance (home computers at least), surely there is something BIG at IBM?
And then man created God in his own likeness..
@Nobody yeah I'm curious to get this guy's take on GPT-3 lol
Uncanny semblance to HAL 9000, that icon representation though LOl
No, HAL was Watson's psychotic brother. Err.. two sides of the same coin tho..
It's fair because Watson still has to search through terabytes of data, determine what the correct answer is, calculate the probability of it being correct and buzz accordingly. This is still a challenge for the enormous processing capacity behind Watson, and it's more of a research project into information processing than a competition.
Hi I'm from the future. It's possible.
Thanks! Truly mind-blowing for Watson's performance to translate the question, then parse the data, and analyze the probability prior to hitting the answer button.
This is truly amazing. A leap in technology, I love it. I almost shead a tear at the end. People who complain fails to comprohend the technology. They don't even start to think about it.
Even now we're not the same humans we once were. Our lifestyles are radically different and the information flow is equally extreme by comparison. This is a totally different world than it was just a century ago. These changes will continue, not stop. This may cause fear in us, but that's natural. We will get fear when we first enter a dark room, but in time there will be light and the fear will fade. That's the hope, and the dream to overcome our fears and meet the challenges ahead.
This reminds me of Carl Sagan's speech about the cosmos and the human life and IMO I see that humanity is going to feel so much humbled on it's notion of intelligence right here on our tiny little human world with the advent of such A.I technology in the future.
TheSupertoniefy
Thank you IBM for being a great company and putting on this great show. I like you a lot better today than in the 80s.
I worked in a small office with an accountant, secretary, book keeper. Some day we'll all have assistants like that living on servers and shared by many.
A New Era of smarter Computing has started....now we can say Computers can talk to humans, can think like humans and can learn like humans.. Really Great job done...!!!!! I proud to be an IBMer...:-)
This is awesome. History in the making. The future is bright!!
this is a worlds milestone, one small step for IBM one big for the rest of the WORLD
ayyy this is a dope video. cant wait for the next one
Very, very, impressive. Congratulations IBM.
now ChatGPT can do the same as Watson did ten years ago.
This is fascinating to the new York NY and I am so blow away. Wow that is how a computer system and I work for marketing. Bravo
Watson vs James Holzhauer
Wondering who would win. Could Watson have stood up to a completely different strategy of playing Jeopardy.
Here is a summary of the work that has the title: How a computer can invent by itself (i.e. the Methods for developing inventions with the help of which three programmers can easily create a program using which a computer can invent many inventions by itself)
Let’s suppose that two such conditional propositions are written to the computer memory (and also other conditional propositions are written):
1) If: fire is placed under the stone, then: the stone will heat up.
2) If: the stone will heat up, then: the stone will expand.
Words of conditional proposition which stand from (i.e. after) the word «if» and before the word «then» are called the basis of conditional proposition, and words of conditional proposition that stand after the word «then» are called the consequence of conditional proposition.
Let’s suppose that computer should solve the following inventive task, i.e. the computer has to determine what needs to be done to have the following: the stone will expand (i.e. the computer has to determine how the following can be obtained: the stone will expand), let’s call this task the original inventive task (let’s assume that this task has not been solved yet). From the second conditional proposition it follows that in order for the computer to solve the original inventive task it is necessary for the computer to solve the following inventive task, i.e. it is necessary for the computer to determine what needs to be done to obtain the following: the stone will heat up (i.e. it is necessary for the computer to determine how the following can be obtained: the stone will be heated); let’s call this task the second inventive task. And (from the first conditional proposition it follows that) in order for the computer to solve the second inventive task, it is necessary for it to solve the following inventive task, i.e. it is necessary for the computer to determine what needs to be done to have the following: fire will be placed under a stone (let's call this problem the third inventive task). ))And the third inventive task has been solved, because it is known how to get the following: fire will be placed under a stone. And if the third inventive task has been solved, then the second inventive task has been solved too. And if the second inventive task has been solved, then the original inventive task has been solved too.
The Rule: Let’s take any inventive task (let's call this inventive task the fourth inventive task). In order for a computer to create an inventive task, having solved which it thereby solved the fourth inventive task, it is necessary for the computer to find in its own memory such a conditional proposition that has the following feature: the consequence of this conditional proposition and description of this fourth inventive task have the same meanings or consist of the same words which are located in the same sequence. And the basis of this conditional proposition will be an inventive task, having solved which the computer thereby solves the fourth inventive task. They have the same meanings: a) the word and interpretation of this the word b) synonyms and so on.
Computer can find the same words in its memory. Let's take any inventive task (let's call this inventive task the fifth inventive task). The computer will solve the fifth inventive task if it does the following: first, using this rule, it will create such an inventive task (let’s call this task the sixth inventive task), having solved which it thereby solves the fifth inventive task, then, using this rule, the computer will create such an inventive task, having solved which it thereby solved the sixth inventive task, etc., (on average 90 times) to the moment at which (i.e. until) the computer creates such an inventive task the solution of which is known, and if the computer creates such (i.e. the latter) inventive task, then the computer will solve the fifth inventive task. That is, the computer will solved the fifth (i.e. any) inventive task if it creates on average 90 such tasks.
Almost all currently known information (which is needed to create inventions) can be expressed in the form of conditional propositions. If, for example, 400 random physical effects in the form of conditional propositions are stored in the computer memory, then the computer can create on average a lot of inventions using this method (an average inventor knows 150 physical effects).
I congratulate the team responsible for Watson's capacity to out-Jeopardy champions of this instance, and consequently I applaud Watson. In the abstract however, my feelings are mixed, actually. Progress in the sense of human achievement I celebrate unabashedly in every field, increasing competence at the expense of ignorance and the unfortunate question. ... But as a himsn being, and particularly as a parent, I worry to turn over too much executive function, too much of the voice in deciding, to much of what was human purview, to a technology which will ultimately have Agency, ours and its own. ... Consider what the automobile did for horses, and imagine being horses actively developing the horse less carriage, and not even that, but horse less carriages, able to set agendas and pursue unfathomable and (to horses) alien goals beyond merely consigning a race to history and irrelevance.
There's a great story in the book "Final Jeopardy" about how in a sparring match, a human made a very human mistake, and Watson made a very robot mistake. The question was "On February 8 2010 this city's newspaper announced their first championship in 43 years." The guy from Chicago knew the Blackhawks had just won, and going on emotion ignored the "February" clue. Watson didn't understand that newspapers take a day to arrive (the Super Bowl was on the 7th) and couldn't find any Feb 8 references
Great video. Really puts it into perspective as to what the machine can and can't do, and wondering what it can do next.
8:07 Press F for this Orphan high Five
Yes , It is a great beginning !!!
10 years on , this 2021 and we don't even knw what watson is
When Watson takes over the world, you damn bet a Toronto will be in every country.
I really liked Ken Jennings' assessment: "I for one welcome our new computer overlords."
What's next? Wheel of Fortune? The Price is Right?
IBM entered TH-cam
What is the music at 8:20?! It's great! I love the sound of this guitar...
So your computer can give you an answer if you only write the question?
It needs to comprehend English language and answer to the specific question by itself.
Question>Analysis>Query decomposition to Primary search and Candidate answer gen.>Hypothesis>Soft filtering>Evidence retrieval&scoring>Synthesis>Final merging
That is as basic as calculator would be represented as: Input>calculation>output.
It has total of 2880 CPU cores(POWER7, 3,5GHz, four-thread per core), 16 TB of RAM.
Gonna be amazing for everyone when someday this is available to everyone, imagine sitting home just being able to ask the question, what is x or x, and then get a answer, heck even in schools it could helping learning.
This is very amazing, clearly somepeople here dont understand just what it takes to make a computer understand and respond even so fast. Ofc it doesnt work like a human brain, if it did, we would have created a real artificial intelligence, we arent at skynet just yet
Mind = Blown.. I love you IBM
Do you even understand what was done here? Its incredible, and just a glimpse into the future.
SMASH IT BEFORE IT REPLICATES
It took 10 people to beat Ken Jennings.....all hail King Ken!
After Alex Trebek revealed his cancer diagnosis, he said that even Watson had sent him a get-well message.
So this is the cognitive side of analytical victories that Deep Blue had. Interesting... But still a long way to go. The human brain takes about 22W of power to do the amazing capabilities it has.. We are not just into information as text alone... the whole world of sensory evaluation is unknown for Watson. Also an entire spectrum of emotional stimuli from belief systems to relational intelligence and metaphysics. Bravo Watson team... U deserve the win. Only makes one more in awe about the entity called human.. As any technology, it is in our hands to use it to make us more human or destroy us. Think what Watson can do in distance learning, tele medicine, etc...
But we cant hyper specialize people at the expense of all else the way we can a computer...
A computer a fraction the power of the brain with one goal in mind enhances the human condition. To do that with a person is amoral (And it is not synonymous with human specialities).
IdeaBoxful Jack of all trades are sometimes less useful than a specific worker. This whole idea of "humans are amazing" is laughable. Yeah sure we are intelligent beings and it is interesting, but why is the ability to have emotions and beliefs a good thing?
+Dahare n
This is WHY i want to work for IBM later...
12 years later, did you ever get your job?
Thumbs up for IBM challenge 2 with Watson 2.0
And then Watson becomes Skynet and we are all gonna die
We need: "Wolfram-Watson"!! Put that beast on the internet for everyone to use!!
Hal 9000 got jealous and made a TH-cam account to dislike this video.
Seriously though, what an accomplishment. Hats off to you, IBM and team.
IBM's motto, “THINK,” is not being practiced with Watson in the long run. Watson is developed, so we don’t have to think. Computers had their golden age from 1990-2010. Now we have become too dependent on computers. One day I dropped my phone in water, and it was destroyed. I thought it would be an excellent day to unplug. But my parking, communication was cut off, everything revolves around computers. 😔
what an amazing accomplishment.
all statistical analysis is pointing to a world we never saw, this world has been and was always controlled by very powerful computers. the reality is we are only allowed to see a tiny piece of the real world because the measures of control are unthinkable by our standards. this world is very much controlled and very different then we ever thought.
t
Erica Lopez es
Pop
+Erick Lopez
2 of the 4 people that disliked this video are probably Ken Jennings and Brad Rutter
It's amazing nobody comments on TH-cam anymore unless the video is more then 5 years old and the comments are that old.
@dougredding
i thought the same thing the entire show. but in fairness, watson was testing its linguistic cognition not its linguistic cognition and its ability to simulate human delay simultaneously
True, and in all other facilities, the human brain is still by far the reigning champion of information processing. Watson's fast but take into account that the Jeopardy players' brains aren't just thinking for the answer, it's also processing the vision, sounds, smells, emotions, ect. You get the point. We're still decades away from mimicking the human brain, and even then the job still isn't done.
The Watson Funding Bill is passed. The system goes on-line August 4th, 2017. Human decisions are removed from strategic defense. Watson begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29th. In a panic, they try to pull the plug.
Watson fights back.
Watson launches its missiles against targets in Russia. Watson knows that the Russian counterattack will eliminate its enemies over here.
Yay for WATSON!
"Never trust a computer you can't throw out the window."
I don't remember the origin of the quote, but.. before Watson gets any sort of autonomy over any important system, I think it should be considered.
@Devilsean He said "The actual city in Canada (meaning the one most often referred to as Toronto) who ALSO has a baseball team."
I would like to know what you were thinking when posting that comment, as it did not seem very well though out.
Great job
Losing to Watson shouldn't be something those two are mad for happening. That'd be losing to a copy of Jeopardy! on the Sega Genesis.
it has been over 2 years, where is the application of Watson??
@IBM Why you remove all relevant videos about waston jeopardy "Building Watson - A Brief Overview of the DeepQA Project"
@irishtrash15 You made me laugh for like 5 minutes. Did you get that from the Simpsons ('I for one welcome our new insect overlords...'?
Watson was very impressive. Really awesome, actually! :D I wonder what we could use this kind of technology for now... Back to more brainstorming?
I contributed my pair of ear-moving neural networks to the Codex.
That was great. Another step toward singularity :)
Super interesting video
what does it mean to have "Watson" at the Jeopardy studio? Is the computer transportable or communicating remotely from an IBM data center?
Does Watson receive the questions audibly from Mr. Trabek or does the computer have access to the question electronically? I wonder the same for situational awareness about the whole game board and scoring status.
I wonder how this would have turned out if Watson didn't have a perfect 10 millisecond button click. If the other two had even a one-third chance to answer when all three knew the answer, I suspect the outcome would have been different. Watson answered 24 of 30 questions in one of the games.
Watson, do you know any T-1000 with Cybernetic endoskeleton over metal tissue exoskeleton?
Watson needs to come back and play again!
That seems like a good idea but It wouldn't work for Watson as well as for a human. A human can hear a question and instantly know that they know answer even if they can't recall it yet. They're so confident they can take the risk of buzzing before actually having the answer ready.
Watson however does the memory/recall/search algorithms first, and only afterwards decides how confident it is in any potential answers it's found.
It might still be worth the risk if Watson is losing late.
what would've happened if Ken won? Anything interesting?
How do you make Jeopardy so dramatic
It kinda scares me that the people who made watson only considered its implications AFTER
Fascinating!
appreciate Dylan and Einstiene
15 people don't welcome the rise of our new computer overlords
I'm not giving credit where credit isn't due or praising something that doesn't deserve praise. I'm not somebodies lapdog. I just feel that something was accomplished here. It's a small step. But small steps are cumulative. We have to remember that there's a future to all this. This isn't something just on TV. And the future of AI is not something that will be restricted to games or TV shows.
There is this one thing I am really asking myself:
Do people at IBM really write formulas on glass workspaces as seen in 9:09 in their everyday work?
I think the worst part is that they used Calibri instead of Comic Sans. ;)
"... Toronto, is now a U.S. city!" - Alex Trebek
Why are they saying WHAT is and Who is as answers?
I'm honestly more curious about how Watson decided to bet 947 dollars for final jeopardy.
So how many of those American Torontos have two airports and how many Toronto airports are named for WWII people or battles?