We all know the server room was a facade, right? I kept waiting for the true behind-the-scenes look at Watson: the room of tanks burgeoning with thousands upon thousands of developing Ken Jennings clones, all neurally intertwined and breathing, thinking, and buzzing as one.
Of course you cordially welcomed our new overlords, Ken. With your genes.
It was super fucking impressive, but contrary to IBM's suggestion that one of the key areas where this sort of open question technology would be useful is in medicine, I'm still pretty sure I don't want my healthcare managed by a set of algorithms that came up with "Toronto" as an answer in the category "U.S. Cities."
Although within another few years of actually refining the algorithms and also tailoring a system to a specific application like medicine, I can possibly see myself changing my mind.
I don't think the goal is to let you walk into a hospital staffed with computer scientists, lay down and give a list of your symptoms to one of them who plugs it into Watson and you get a diagnosis.
It's an information tool, and a damn good one. Doctors who might spend days poring through books looking for some mysterious illness could instead have Watson give them some clues in a few seconds. Then they could weigh how correct or incorrect they may be and go from there.
Everybody in an advanced profession is confronted with something they know little about. This sort of thing would just make information easier and faster to obtain.
All this talk about being in hospitals makes me want an episode of House with special guest Watson.
I once had a dream about a House episode sort of like that. It was about a computer programmer who develops an extremely accurate computer-based diagnostic tool (I think it was called iHouse in the dream). Unfortunately, the computer programmer comes down with some mysterious illness that iHouse is unable to diagnose. House treats him because its mysterious and, as usual, eventually figures out what it was. The end of the episode is House deducing that the programmer got himself sick with an illness that iHouse could't diagnose because he was noticing that many hospitals were replacing real doctors with iHouse and he wanted to show them that iHouse was just a tool and you still needed a human doctor.
It was super fucking impressive, but contrary to IBM's suggestion that one of the key areas where this sort of open question technology would be useful is in medicine, I'm still pretty sure I don't want my healthcare managed by a set of algorithms that came up with "Toronto" as an answer in the category "U.S. Cities."
Although within another few years of actually refining the algorithms and also tailoring a system to a specific application like medicine, I can possibly see myself changing my mind.
I don't think the goal is to let you walk into a hospital staffed with computer scientists, lay down and give a list of your symptoms to one of them who plugs it into Watson and you get a diagnosis.
That's not what I'd be leery about. My mother being a medical researcher, I've heard more than a few times about doctors who come up with one explanation that fits a set of symptoms and then stop looking for a second, even when the first explanation ends up being wrong. It's the root cause of the existence of the phrase, "can I get a second opinion?" So I'm inherently skeptical of the idea of prejudicing a doctor's medical opinions by simply giving him the answer rather than encouraging him to read that pile of books.
It's actually no so much that I'm unimpressed by the computer as that I'm frequently unimpressed by human psychology.
It was super fucking impressive, but contrary to IBM's suggestion that one of the key areas where this sort of open question technology would be useful is in medicine, I'm still pretty sure I don't want my healthcare managed by a set of algorithms that came up with "Toronto" as an answer in the category "U.S. Cities."
Although within another few years of actually refining the algorithms and also tailoring a system to a specific application like medicine, I can possibly see myself changing my mind.
I don't think the goal is to let you walk into a hospital staffed with computer scientists, lay down and give a list of your symptoms to one of them who plugs it into Watson and you get a diagnosis.
That's not what I'd be leery about. My mother being a medical researcher, I've heard more than a few times about doctors who come up with one explanation that fits a set of symptoms and then stop looking for a second, even when the first explanation ends up being wrong. It's the root cause of the existence of the phrase, "can I get a second opinion?" So I'm inherently skeptical of the idea of prejudicing a doctor's medical opinions by simply giving him the answer rather than encouraging him to read that pile of books.
It's actually no so much that I'm unimpressed by the computer as that I'm frequently unimpressed by human psychology.
Pre-Watson:
"Here's the diagnosis"
"I'm getting a second opinion"
Post-Watson:
"I'm going to take this list of possibilities to another doctor and see what they think"
With one you're depending 100% on the knowledge of those two doctors. But with a list you can insist that both doctors consider a few possibilities with the knowledge that they exist.
All this talk about being in hospitals makes me want an episode of House with special guest Watson.
I once had a dream about a House episode sort of like that. It was about a computer programmer who develops an extremely accurate computer-based diagnostic tool (I think it was called iHouse in the dream). Unfortunately, the computer programmer comes down with some mysterious illness that iHouse is unable to diagnose. House treats him because its mysterious and, as usual, eventually figures out what it was. The end of the episode is House deducing that the programmer got himself sick with an illness that iHouse could't diagnose because he was noticing that many hospitals were replacing real doctors with iHouse and he wanted to show them that iHouse was just a tool and you still needed a human doctor.
It was super fucking impressive, but contrary to IBM's suggestion that one of the key areas where this sort of open question technology would be useful is in medicine, I'm still pretty sure I don't want my healthcare managed by a set of algorithms that came up with "Toronto" as an answer in the category "U.S. Cities."
Although within another few years of actually refining the algorithms and also tailoring a system to a specific application like medicine, I can possibly see myself changing my mind.
I don't think the goal is to let you walk into a hospital staffed with computer scientists, lay down and give a list of your symptoms to one of them who plugs it into Watson and you get a diagnosis.
It's an information tool, and a damn good one. Doctors who might spend days poring through books looking for some mysterious illness could instead have Watson give them some clues in a few seconds. Then they could weigh how correct or incorrect they may be and go from there.
Everybody in an advanced profession is confronted with something they know little about. This sort of thing would just make information easier and faster to obtain.
Things like this have also been developed for doing literature surveys. The volume of published research out there is simply staggering and a lot of research, or at least background development for research, requires the analysis and collection of it together to identify trends etc.
Doing this by hand takes a very long time, and that problem is only increasing. Something like Watson, capable of identifying concepts and relating them, would be (and has been - I read about something like this a few years back specifically more assisting in medical research) incredibly useful.
Being able to persistently do meta-analyses on research papers would be amazing.
Short of it is they are taking two years to do re-vamping on Watson before putting him in two test hospitals; with most of the revamp consisting of being able to understand medical speak and being able to come up with multiple hypotheses.
We all know the server room was a facade, right? I kept waiting for the true behind-the-scenes look at Watson: the room of tanks burgeoning with thousands upon thousands of developing Ken Jennings clones, all neurally intertwined and breathing, thinking, and buzzing as one.
Of course you cordially welcomed our new overlords, Ken. With your genes.
I thought that off camera, there was a curtain hiding a wizard...
Also, is it me or did Jennings toss the buzzer onto the ground behind Trebek during the credits?
It was super fucking impressive, but contrary to IBM's suggestion that one of the key areas where this sort of open question technology would be useful is in medicine, I'm still pretty sure I don't want my healthcare managed by a set of algorithms that came up with "Toronto" as an answer in the category "U.S. Cities."
Although within another few years of actually refining the algorithms and also tailoring a system to a specific application like medicine, I can possibly see myself changing my mind.
I don't think the goal is to let you walk into a hospital staffed with computer scientists, lay down and give a list of your symptoms to one of them who plugs it into Watson and you get a diagnosis.
That's not what I'd be leery about. My mother being a medical researcher, I've heard more than a few times about doctors who come up with one explanation that fits a set of symptoms and then stop looking for a second, even when the first explanation ends up being wrong. It's the root cause of the existence of the phrase, "can I get a second opinion?" So I'm inherently skeptical of the idea of prejudicing a doctor's medical opinions by simply giving him the answer rather than encouraging him to read that pile of books.
It's actually no so much that I'm unimpressed by the computer as that I'm frequently unimpressed by human psychology.
Pre-Watson:
"Here's the diagnosis"
"I'm getting a second opinion"
Post-Watson:
"I'm going to take this list of possibilities to another doctor and see what they think"
With one you're depending 100% on the knowledge of those two doctors. But with a list you can insist that both doctors consider a few possibilities with the knowledge that they exist.
Yeah see I don't think that's at all better. Now you're just prejudicing two doctors' opinions instead of one. If you ever happen to find yourself in a position that you want a second opinion, don't tell them what the first doctor said the possibilities are -- give him a copy of all your test results and tell him your symptoms without telling him what the first diagnosis was. That way you can get an independent consult and see whether or not the two diagnoses coincide; otherwise you run the risk that the doctor will use the original diagnosis as the basis for his own conclusions as opposed to your symptoms.
EDIT: One possibility I've just considered is that maybe Watson, M.D. could keep his possible diagnoses to himself (itself, I suppose?) and instead suggest a couple of different tests which would help the doctor in making a diagnosis on his own. While spitting out a literature survey for the doctor to read through while his lab techs run the tests.
I have to imagine that Doctor's get really frustrated when a Patient cites Google as to what their diagnosis might be, or what a better treatment might be than what the Doctor suggests. I can picture now, some tool walking into a Doctor's office, telling the Doctor that he has all the Symptoms of Lupus and needs some experimental drug. The Doctor assures them it isn't Lupus, but the Patient insists it is.
The Doctor plugs the Patients symptoms into Watson, and Watson spits out 20 possibilities, with Lupus ringing in at possibility #1 because of the millions of people searching Google thinking they have Lupus. Doctor says "Fuck this, I quit. I'm gonna be a professional Poker Player." He moves to Las Vegas and finds out that all the Pro Poker Players are Robots. He would rather be dead than live in a world ruled by computers, and so ends it all.
My Puny Human Brain Jeopardy! genius Ken Jennings on what it's like to play against a supercomputer.
When I was selected as one of the two human players to be pitted against IBM's "Watson" supercomputer in a special man-vs.-machine Jeopardy! exhibition match, I felt honored, even heroic. I envisioned myself as the Great Carbon-Based Hope against a new generation of thinking machines—which, if Hollywood is to believed, will inevitably run amok, build unstoppable robot shells, and destroy us all. But at IBM's Thomas J. Watson Research Lab, an Eero Saarinen-designed fortress in the snowy wilds of New York's Westchester County, where the shows taped last month, I wasn't the hero at all. I was the villain.
This was to be an away game for humanity, I realized as I walked onto the slightly-smaller-than-regulation Jeopardy! set that had been mocked up in the building's main auditorium. In the middle of the floor was a huge image of Watson's on-camera avatar, a glowing blue ball crisscrossed by "threads" of thought—42 threads, to be precise, an in-joke for Douglas Adams fans. The stands were full of hopeful IBM programmers and executives, whispering excitedly and pumping their fists every time their digital darling nailed a question. A Watson loss would be invigorating for Luddites and computer-phobes everywhere, but bad news for IBM shareholders.
I'm glad he's apparently taking this better than Kasparov did.
I don't get why Kasparov freaked out. I mean, chess is pretty obviously finite and would eventually fall to a big enough difference engine.
Well I think there is pretty much a consensus among chess players that computers (especially at that time) had a certain mode of thinking. They were thus incapable of certain strategic maneuvers and all the games Kasparov and Deep Blue played this held true. But then suddenly in a freak move Deep Blue broke these rules and thus made a move "only a human could make". And computer have only very recently been able to use the same strategic moves as humans, which is why fairly good Humans with a pretty good chess computer (that can help them just crunch numbers) beat both the best human players and the best computer players.
So it is pretty obvious that IBM cheated. Of course this has little bearing on computer eventually beating humans, today it is not even close, but cheating is cheating and I can see why he was upset, and still is.
fjafjan on
Yepp, THE Fjafjan (who's THE fjafjan?)
- "Proving once again the deadliest animal of all ... is the Zoo Keeper" - Philip J Fry
I'm glad he's apparently taking this better than Kasparov did.
I don't get why Kasparov freaked out. I mean, chess is pretty obviously finite and would eventually fall to a big enough difference engine.
Well I think there is pretty much a consensus among chess players that computers (especially at that time) had a certain mode of thinking. They were thus incapable of certain strategic maneuvers and all the games Kasparov and Deep Blue played this held true. But then suddenly in a freak move Deep Blue broke these rules and thus made a move "only a human could make". And computer have only very recently been able to use the same strategic moves as humans, which is why fairly good Humans with a pretty good chess computer (that can help them just crunch numbers) beat both the best human players and the best computer players.
So it is pretty obvious that IBM cheated. Of course this has little bearing on computer eventually beating humans, today it is not even close, but cheating is cheating and I can see why he was upset, and still is.
I would be much more skeptical of Kasparov's accusations if IBM hadn't refused to provide Deep Blue's log files and dismantled the computer. It no longer really matters, anyway, computers have shown themselves to be unequivocally better at chess than humans since then.
I couldn't find a contemporary news article that wasn't pay-per-view, but he basically accused IBM of cheating -- he suggested that the computer's creativity meant that a human was feeding the computer moves, which is why he demanded to see the log files.
He also complained that he hadn't been allowed to study any of Deep Blue's games where as IBM had fed many of Kasparov's games into Deep Blue, which he claimed created an unnatural disadvantage.
Re: Watson, the natural language recognition bit can stand as brilliant on its own merits; but it also seemed to have an uncanny ability to get inside of whatever determined where to find the double-jeopardies (correct me if I'm wrong, but Watson found five out of six of them, right?) which undoubtedly influenced the totals going into final jeopardy. I could see the other players getting annoyed about that if they hadn't stopped to remember that it's only a game.
but it also seemed to have an uncanny ability to get inside of whatever determined where to find the double-jeopardies (correct me if I'm wrong, but Watson found five out of six of them, right?) which undoubtedly influenced the totals going into final jeopardy.
I think that just happened because he got to choose the clues 90% of the time.
but it also seemed to have an uncanny ability to get inside of whatever determined where to find the double-jeopardies (correct me if I'm wrong, but Watson found five out of six of them, right?) which undoubtedly influenced the totals going into final jeopardy.
I think that just happened because he got to choose the clues 90% of the time.
The first one was a real "Oh come ON!" moment, but for the most part I agree. He found them because he was doing the picking almost every time.
But I'd be willing to bet that IBM fed thousands of games of jeopardy into Watson and it determined, statistically, that the far left category $800 square holds the daily double most often.
I say this because if that square was still on the board and Watson had control, it picked that square.
The probabilities of double jeopardy is something guys like Rutter and Jennings are also aware of, maybe not to the exact percentage but they know where they should be looking. And Jennings did hit one even though he got to pick a minority of the time by far.
Soooo... How close are we now to having a Commander Data? I saw a quick blurb on the news that Japan is sending a human-like robot to the space station. It looks like those weird-robots they've been showing in the Adult Swim bumpers recently. Uncanny valley aside, if something like Watson could be sorta combined with this, isn't this a few steps closer to a Data?
The probabilities of double jeopardy is something guys like Rutter and Jennings are also aware of, maybe not to the exact percentage but they know where they should be looking. And Jennings did hit one even though he got to pick a minority of the time by far.
The weird thing about that? He probably wouldn't have gotten a chance to if there hadn't been a bug in Watson's programming.
It's a known "thing" that the daily doubles tend to be on one specific area of the board, because they are not placed randomly. Jennings even mentioned that he was sure the last daily double that Watson got was in the category, but had to decide which one of the two clues left had it and ended up picking the wrong.
So IBM definitely programmed his Jeopardy algorithms to seek out the daily doubles.
It's a known "thing" that the daily doubles tend to be on one specific area of the board, because they are not placed randomly. Jennings even mentioned that he was sure the last daily double that Watson got was in the category, but had to decide which one of the two clues left had it and ended up picking the wrong.
So IBM definitely programmed his Jeopardy algorithms to seek out the daily doubles.
Yup, they did. They did a lot of game-related programming to make sure Watson "knew what it was doing." Wagering theory, DJ placement, etc. The funny part is, from what I can tell, they could make just a couple minor tweaks in the game-related AI and make Watson even more unbeatable, given what they saw in these games.
I have to imagine that Doctor's get really frustrated when a Patient cites Google as to what their diagnosis might be, or what a better treatment might be than what the Doctor suggests. I can picture now, some tool walking into a Doctor's office, telling the Doctor that he has all the Symptoms of Lupus and needs some experimental drug. The Doctor assures them it isn't Lupus, but the Patient insists it is.
The Doctor plugs the Patients symptoms into Watson, and Watson spits out 20 possibilities, with Lupus ringing in at possibility #1 because of the millions of people searching Google thinking they have Lupus. Doctor says "Fuck this, I quit. I'm gonna be a professional Poker Player." He moves to Las Vegas and finds out that all the Pro Poker Players are Robots. He would rather be dead than live in a world ruled by computers, and so ends it all.
getting something like this in the US will just cause a lot of abuse and annoyance.
Getting this to a third world country however can save a massive amount of lives.
DanHibiki on
0
Options
GreasyKidsStuffMOMMM!ROAST BEEF WANTS TO KISS GIRLS ON THE TITTIES!Registered Userregular
edited February 2011
From that Ken Jennings response article, this got a huge laugh out of me
Watson has lots in common with a top-ranked human Jeopardy! player: It's very smart, very fast, speaks in an uneven monotone, and has never known the touch of a woman.
I'm sad I missed out on watching this, but I think it might've been PVR'd. I'll have to look.
I have to imagine that Doctor's get really frustrated when a Patient cites Google as to what their diagnosis might be, or what a better treatment might be than what the Doctor suggests. I can picture now, some tool walking into a Doctor's office, telling the Doctor that he has all the Symptoms of Lupus and needs some experimental drug. The Doctor assures them it isn't Lupus, but the Patient insists it is.
The Doctor plugs the Patients symptoms into Watson, and Watson spits out 20 possibilities, with Lupus ringing in at possibility #1 because of the millions of people searching Google thinking they have Lupus. Doctor says "Fuck this, I quit. I'm gonna be a professional Poker Player." He moves to Las Vegas and finds out that all the Pro Poker Players are Robots. He would rather be dead than live in a world ruled by computers, and so ends it all.
getting something like this in the US will just cause a lot of abuse and annoyance.
Getting this to a third world country however can save a massive amount of lives.
One of the stories I tell in "Blink" is about the Emergency Room doctors at Cook County Hospital in Chicago. That's the big public hospital in Chicago, and a few years ago they changed the way they diagnosed heart attacks. They instructed their doctors to gather less information on their patients: they encouraged them to zero in on just a few critical pieces of information about patients suffering from chest pain--like blood pressure and the ECG--while ignoring everything else, like the patient's age and weight and medical history. And what happened? Cook County is now one of the best places in the United States at diagnosing chest pain.
The book, Blink, is an interesting enough read.
The section the above is talking about is really good, though: turns out doctors are pretty shitty at listening to patients symptoms and coming up with an appropriate way to treat them when it comes to chest pain. They're biased towards treating patients based on the number of symptoms listed, applying an equal sort of ranking to each piece of knowledge (humans do this in general, if I can think of the term for it I'll link to an article about it) when what's actually important is recognizing key symptoms that indicate you'd better get the patient taken care of right now.
The book illustrates this quite clearly by giving two examples of people with chest pain symptoms and their responses to their doctor's questions. And you'd swear, looking at it, that one of them needs to get into ER right the hell now.
Turns out, nope: sure, he's going to have a heart attack some day, but it's probably not today. The docs weren't really happy about what amounts to running through a script of what information they needed to gather and how they had to evaluate it, but it's made a hell of a difference.
I for one welcome our new robot doctors. Provided they're scripted properly .
Ego on
Erik
0
Options
chiasaur11Never doubt a raccoon.Do you think it's trademarked?Registered Userregular
Posts
Just... think about that one for a second.
Of course you cordially welcomed our new overlords, Ken. With your genes.
I don't think the goal is to let you walk into a hospital staffed with computer scientists, lay down and give a list of your symptoms to one of them who plugs it into Watson and you get a diagnosis.
It's an information tool, and a damn good one. Doctors who might spend days poring through books looking for some mysterious illness could instead have Watson give them some clues in a few seconds. Then they could weigh how correct or incorrect they may be and go from there.
Everybody in an advanced profession is confronted with something they know little about. This sort of thing would just make information easier and faster to obtain.
I once had a dream about a House episode sort of like that. It was about a computer programmer who develops an extremely accurate computer-based diagnostic tool (I think it was called iHouse in the dream). Unfortunately, the computer programmer comes down with some mysterious illness that iHouse is unable to diagnose. House treats him because its mysterious and, as usual, eventually figures out what it was. The end of the episode is House deducing that the programmer got himself sick with an illness that iHouse could't diagnose because he was noticing that many hospitals were replacing real doctors with iHouse and he wanted to show them that iHouse was just a tool and you still needed a human doctor.
That's not what I'd be leery about. My mother being a medical researcher, I've heard more than a few times about doctors who come up with one explanation that fits a set of symptoms and then stop looking for a second, even when the first explanation ends up being wrong. It's the root cause of the existence of the phrase, "can I get a second opinion?" So I'm inherently skeptical of the idea of prejudicing a doctor's medical opinions by simply giving him the answer rather than encouraging him to read that pile of books.
It's actually no so much that I'm unimpressed by the computer as that I'm frequently unimpressed by human psychology.
Pre-Watson:
"Here's the diagnosis"
"I'm getting a second opinion"
Post-Watson:
"I'm going to take this list of possibilities to another doctor and see what they think"
With one you're depending 100% on the knowledge of those two doctors. But with a list you can insist that both doctors consider a few possibilities with the knowledge that they exist.
I would watch House if it had robots.
3DS FC: 5343-7720-0490
Things like this have also been developed for doing literature surveys. The volume of published research out there is simply staggering and a lot of research, or at least background development for research, requires the analysis and collection of it together to identify trends etc.
Doing this by hand takes a very long time, and that problem is only increasing. Something like Watson, capable of identifying concepts and relating them, would be (and has been - I read about something like this a few years back specifically more assisting in medical research) incredibly useful.
Being able to persistently do meta-analyses on research papers would be amazing.
http://tech.ca.msn.com/canadianpress-article.aspx?cp-documentid=27708421
Short of it is they are taking two years to do re-vamping on Watson before putting him in two test hospitals; with most of the revamp consisting of being able to understand medical speak and being able to come up with multiple hypotheses.
I thought that off camera, there was a curtain hiding a wizard...
Also, is it me or did Jennings toss the buzzer onto the ground behind Trebek during the credits?
Yeah see I don't think that's at all better. Now you're just prejudicing two doctors' opinions instead of one. If you ever happen to find yourself in a position that you want a second opinion, don't tell them what the first doctor said the possibilities are -- give him a copy of all your test results and tell him your symptoms without telling him what the first diagnosis was. That way you can get an independent consult and see whether or not the two diagnoses coincide; otherwise you run the risk that the doctor will use the original diagnosis as the basis for his own conclusions as opposed to your symptoms.
EDIT: One possibility I've just considered is that maybe Watson, M.D. could keep his possible diagnoses to himself (itself, I suppose?) and instead suggest a couple of different tests which would help the doctor in making a diagnosis on his own. While spitting out a literature survey for the doctor to read through while his lab techs run the tests.
The Doctor plugs the Patients symptoms into Watson, and Watson spits out 20 possibilities, with Lupus ringing in at possibility #1 because of the millions of people searching Google thinking they have Lupus. Doctor says "Fuck this, I quit. I'm gonna be a professional Poker Player." He moves to Las Vegas and finds out that all the Pro Poker Players are Robots. He would rather be dead than live in a world ruled by computers, and so ends it all.
http://www.slate.com/id/2284721/
My Puny Human Brain
Jeopardy! genius Ken Jennings on what it's like to play against a supercomputer.
When I was selected as one of the two human players to be pitted against IBM's "Watson" supercomputer in a special man-vs.-machine Jeopardy! exhibition match, I felt honored, even heroic. I envisioned myself as the Great Carbon-Based Hope against a new generation of thinking machines—which, if Hollywood is to believed, will inevitably run amok, build unstoppable robot shells, and destroy us all. But at IBM's Thomas J. Watson Research Lab, an Eero Saarinen-designed fortress in the snowy wilds of New York's Westchester County, where the shows taped last month, I wasn't the hero at all. I was the villain.
This was to be an away game for humanity, I realized as I walked onto the slightly-smaller-than-regulation Jeopardy! set that had been mocked up in the building's main auditorium. In the middle of the floor was a huge image of Watson's on-camera avatar, a glowing blue ball crisscrossed by "threads" of thought—42 threads, to be precise, an in-joke for Douglas Adams fans. The stands were full of hopeful IBM programmers and executives, whispering excitedly and pumping their fists every time their digital darling nailed a question. A Watson loss would be invigorating for Luddites and computer-phobes everywhere, but bad news for IBM shareholders.
I don't get why Kasparov freaked out. I mean, chess is pretty obviously finite and would eventually fall to a big enough difference engine.
Well I think there is pretty much a consensus among chess players that computers (especially at that time) had a certain mode of thinking. They were thus incapable of certain strategic maneuvers and all the games Kasparov and Deep Blue played this held true. But then suddenly in a freak move Deep Blue broke these rules and thus made a move "only a human could make". And computer have only very recently been able to use the same strategic moves as humans, which is why fairly good Humans with a pretty good chess computer (that can help them just crunch numbers) beat both the best human players and the best computer players.
So it is pretty obvious that IBM cheated. Of course this has little bearing on computer eventually beating humans, today it is not even close, but cheating is cheating and I can see why he was upset, and still is.
- "Proving once again the deadliest animal of all ... is the Zoo Keeper" - Philip J Fry
I would be much more skeptical of Kasparov's accusations if IBM hadn't refused to provide Deep Blue's log files and dismantled the computer. It no longer really matters, anyway, computers have shown themselves to be unequivocally better at chess than humans since then.
I couldn't find a contemporary news article that wasn't pay-per-view, but he basically accused IBM of cheating -- he suggested that the computer's creativity meant that a human was feeding the computer moves, which is why he demanded to see the log files.
He also complained that he hadn't been allowed to study any of Deep Blue's games where as IBM had fed many of Kasparov's games into Deep Blue, which he claimed created an unnatural disadvantage.
Re: Watson, the natural language recognition bit can stand as brilliant on its own merits; but it also seemed to have an uncanny ability to get inside of whatever determined where to find the double-jeopardies (correct me if I'm wrong, but Watson found five out of six of them, right?) which undoubtedly influenced the totals going into final jeopardy. I could see the other players getting annoyed about that if they hadn't stopped to remember that it's only a game.
I think that just happened because he got to choose the clues 90% of the time.
The first one was a real "Oh come ON!" moment, but for the most part I agree. He found them because he was doing the picking almost every time.
But I'd be willing to bet that IBM fed thousands of games of jeopardy into Watson and it determined, statistically, that the far left category $800 square holds the daily double most often.
I say this because if that square was still on the board and Watson had control, it picked that square.
twitch.tv/Taramoor
@TaramoorPlays
Taramoor on Youtube
The breakthrough here is only marginally related to making conscious machine intelligence.
The weird thing about that? He probably wouldn't have gotten a chance to if there hadn't been a bug in Watson's programming.
I didn't know this was possible
So IBM definitely programmed his Jeopardy algorithms to seek out the daily doubles.
Yup, they did. They did a lot of game-related programming to make sure Watson "knew what it was doing." Wagering theory, DJ placement, etc. The funny part is, from what I can tell, they could make just a couple minor tweaks in the game-related AI and make Watson even more unbeatable, given what they saw in these games.
getting something like this in the US will just cause a lot of abuse and annoyance.
Getting this to a third world country however can save a massive amount of lives.
I'm sad I missed out on watching this, but I think it might've been PVR'd. I'll have to look.
The book, Blink, is an interesting enough read.
The section the above is talking about is really good, though: turns out doctors are pretty shitty at listening to patients symptoms and coming up with an appropriate way to treat them when it comes to chest pain. They're biased towards treating patients based on the number of symptoms listed, applying an equal sort of ranking to each piece of knowledge (humans do this in general, if I can think of the term for it I'll link to an article about it) when what's actually important is recognizing key symptoms that indicate you'd better get the patient taken care of right now.
The book illustrates this quite clearly by giving two examples of people with chest pain symptoms and their responses to their doctor's questions. And you'd swear, looking at it, that one of them needs to get into ER right the hell now.
Turns out, nope: sure, he's going to have a heart attack some day, but it's probably not today. The docs weren't really happy about what amounts to running through a script of what information they needed to gather and how they had to evaluate it, but it's made a hell of a difference.
I for one welcome our new robot doctors. Provided they're scripted properly
Heh.
It's cute how you assume it's that innocent.
We should shove Watson onto Deimos and shoot it out of the solar system. To be safe.
Why I fear the ocean.
Isn't that how the plot of Doom began?