Seems to me that we're playing a game that isn't that hard to win at.
Make no doubt, it is truly amazing. The patterns that must be recognized to answer even the simplest question correct are extremely complex. The real problem is understanding the question AND putting it into the proper perspective of the category, then build and execute the proper queries to get the possible right answers. Then it needs to rank those answers, determine if the certainty is high enough to make an answer and then buzz in..
For example answering a simple question like: "When were Oreo cookies and crossword puzzles invented?" wouldn't be hard because the question is straightforward you can easily determine what it is asking by processing basic linguistic rules in a deterministic fashion. The you simply need to reference any number of databases to find the answer.
Change that to a Jeopardy Style "answer" in the category of (Name the Decade)and you get: "The first modern crossword puzzle is published & Oreo cookies are introduced" and things get more complex.
First you need to determine that the what the category is asking for, which in this case isn't that hard, but at any rate is still very important or you might be picking an answer that is totally inappropriate (Like answering Toronto as the name of a US city). Then you need to parse the question to get the important details.
Bottom line is, 5 seconds with Google and any 12 year old could tell you that they came out in 1912 but Watson got it wrong on Monday.
__________________
"Wake up, Luigi! The only time plumbers sleep on the job is when we're working by the hour."
It's unfortunate that Watson had no auditory sensors or visual receptors.
I'm finding his (its?) skill at Jeopardy to be quite fascinating. But I would be blown away if Trebek could carry on a conversation with it like he does with the other contestants.
Sadly, something like that is still a long ways off.
Make no doubt, it is truly amazing. The patterns that must be recognized to answer even the simplest question correct are extremely complex. The real problem is understanding the question AND putting it into the proper perspective of the category, then build and execute the proper queries to get the possible right answers. Then it needs to rank those answers, determine if the certainty is high enough to make an answer and then buzz in..
For example answering a simple question like: "When were Oreo cookies and crossword puzzles invented?" wouldn't be hard because the question is straightforward you can easily determine what it is asking by processing basic linguistic rules in a deterministic fashion. The you simply need to reference any number of databases to find the answer.
Change that to a Jeopardy Style "answer" in the category of (Name the Decade)and you get: "The first modern crossword puzzle is published & Oreo cookies are introduced" and things get more complex.
First you need to determine that the what the category is asking for, which in this case isn't that hard, but at any rate is still very important or you might be picking an answer that is totally inappropriate (Like answering Toronto as the name of a US city). Then you need to parse the question to get the important details.
Bottom line is, 5 seconds with Google and any 12 year old could tell you that they came out in 1912 but Watson got it wrong on Monday.
Hey, I'm not just being a rube, and saying "hey what's the DEAAAL with computers anyway, they ain't all that good," (see, I'm not even good at typing with a rube accent), I'm just saying that perhaps all we have done in creating a Watson is making a more efficient calculator, and not something that comes even close to human behaviour or intelligence.
I'm not sure what you're trying to say here, espeically by linking to that wikipedia article, but yes it is amazing.
What they've done is build a computer that can recognize natural language including slang, puns, and other things that even other humans have trouble with, if they are new to the language, and have it drawing conclusions from that. That's something that no one has been able to do before, and it's a pretty big step forward.
And yes, that game really is that hard to win at. Being able to piece together the right answer based on clues, not search parameters, is a pretty difficult thing to do.
If it's as easy as you claim it is, then why has no one done it before, and why did it take some of the best engineers at one of the largest computer companies in the world 3 years to accomplish it after they came up with the idea?
Again, I do not doubt the technical prowess of the machine or the technical knowledge of the creators, I just doubt the miraculous potential of such a machine. Watson is an enormously powerful calculator able to simulate relatively limited human interaction.
I worry about two things, right off the bat.
A) That machines like Watson, and his future ilk, will begin to simulate human interaction, such as the machine therapists that Carl Sagan proposed as an answer to human therapy problems.
B) The downward mimesis of machine language by people. Nietzsche and Heiddeger are excellent critics on this, but you can go back to Plato. Basically, language is important, human nature is far more plastic than we want to think, and the cultural capacity to carry language is the most important thing we have to create and understand our world.
Machines are coded in logical terms, series of instructions which they execute. Human conversations are nothing like this, they are organic, feeding off mutual understanding, instinct, and subtle signals.
Watson IS amazing, but all I see is a real neat freak-show.
Hey, I'm not just being a rube, and saying "hey what's the DEAAAL with computers anyway, they ain't all that good," (see, I'm not even good at typing with a rube accent), I'm just saying that perhaps all we have done in creating a Watson is making a more efficient calculator, and not something that comes even close to human behaviour or intelligence.
It really is just a bigger calculator.
__________________
"Wake up, Luigi! The only time plumbers sleep on the job is when we're working by the hour."
Watson had a big advantage in that he played against 2 human opponants. The questions Watson took longer to ring in on were split between Rutter and Jennings.
If there were two Watsons playing and just Jennings the human probably wins. (assuming the Watsons are roughly equal). So a one on one challange would be more fair. I am impressed with what Watson can do and I think the medical diagnosis application is definately something that it would do well.
Instead of Watson though they should have named him Multivac.
I don't buy it at all that Watson doesn't have a massive advantage when it comes to buzzing in first. (2nd article)
In fact I'd say the only reason Watson won is because of the buzzer speed. It is impressive that Watson can get the answers at all but since all three contestants probably know 90% of the right answers the buzzer is the only difference.
For IBM to claim otherwise is BS. You could even see the Ken and Brad, especially Ken, get frustrated because Watson is so lightning fast.
The Following User Says Thank You to Cecil Terwilliger For This Useful Post:
Location: In my office, at the Ministry of Awesome!
Exp:
Quote:
Originally Posted by Cecil Terwilliger
I don't buy it at all that Watson doesn't have a massive advantage when it comes to buzzing in first. (2nd article)
In fact I'd say the only reason Watson won is because of the buzzer speed. It is impressive that Watson can get the answers at all but since all three contestants probably know 90% of the right answers the buzzer is the only difference.
For IBM to claim otherwise is BS. You could even see the Ken and Brad, especially Ken, get frustrated because Watson is so lightning fast.
Again, not the point.
Complaining that Watson has a buzzer advantage is completely moot. Of course a machine is going to have an advantage at physical tasks. Up untill this point though, humans have always had a MASSIVE advantage when it came to the other part of the game, and it looks as though with Watson, that advantabe is now gone.
__________________
THE SHANTZ WILL RISE AGAIN. <-----Check the Badge bitches. You want some Awesome, you come to me!
I don't buy it at all that Watson doesn't have a massive advantage when it comes to buzzing in first. (2nd article)
In fact I'd say the only reason Watson won is because of the buzzer speed. It is impressive that Watson can get the answers at all but since all three contestants probably know 90% of the right answers the buzzer is the only difference.
For IBM to claim otherwise is BS. You could even see the Ken and Brad, especially Ken, get frustrated because Watson is so lightning fast.
I love their explanation too. "oh well a human can randomly take a stab at when they think the light is going to come on vs Watson has to wait for it to actually happen then can process it and click faster then humanly possible. But only if one of the humans didn't fluke it out first."
You know the more I read about this the more it reinforced to me how far off computers still really are from being able to come close to replicating something even remotely similar to a human brain(probably for the better anyways though )
You know the more I read about this the more it reinforced to me how far off computers still really are from being able to come close to replicating something even remotely similar to a human brain(probably for the better anyways though )
When it takes tens of millions of dollars and a computer the size of 10 fridge boxes, and several years of labour for a software development team just to compete on Jeopardy!, and then have it still miss some questions that are totally obvious to any human, yeah I would agree with you.
The logic is amazing, but it shows how much the human brain can do, so much of which we never think about.
I don't buy it at all that Watson doesn't have a massive advantage when it comes to buzzing in first. (2nd article)
In fact I'd say the only reason Watson won is because of the buzzer speed. It is impressive that Watson can get the answers at all but since all three contestants probably know 90% of the right answers the buzzer is the only difference.
For IBM to claim otherwise is BS. You could even see the Ken and Brad, especially Ken, get frustrated because Watson is so lightning fast.
You are giving far too little credit to the computer figuring out the actual answers to the questions. Buzzer speed is nothing if it doesn't know the answer. That's the hard part, not the buzzer.
<-----Check the Badge bitches. You want some Awesome, you come to me!
)
