EPTIBM Watson Computer Beats Jeopardy Pros - Page 9
| Forum Index > General Forum |
|
lixlix
United States482 Posts
| ||
|
Megaman703
Canada688 Posts
I don't think the rules should be changed for this challenge, but I can definitely appreciate the opinion that this has basically turned into a game of Watson vs Jeopardy Clues, not Man vs Machine, and that makes it uninteresting for the viewer. | ||
|
Gov
United States121 Posts
On February 16 2011 14:19 Megaman703 wrote: It's kind of ironic to see Ken Jennings of all people getting frustrated at a competitor out speeding him to the buzzer. I don't think the rules should be changed for this challenge, but I can definitely appreciate the opinion that this has basically turned into a game of Watson vs Jeopardy Clues, not Man vs Machine, and that makes it uninteresting for the viewer. as a computer scientist i find it fascinating, whether it is a close game or not | ||
|
Megaman703
Canada688 Posts
| ||
|
29 fps
United States5725 Posts
now... let's have 3 comps face each other. | ||
|
semantics
10040 Posts
On February 16 2011 13:37 lixlix wrote: 200 Power7s is a lot of computing power. so much so that it will probably be a long while before we see this on a consumer level. | ||
|
Froadac
United States6733 Posts
It's more like machine vs question. If it picks up on it super fast, you're just sorta screwed. It'll be interesting when its' machien vs machine lol. :DD | ||
|
lixlix
United States482 Posts
On February 16 2011 14:59 semantics wrote: so much so that it will probably be a long while before we see this on a consumer level. nah, you can have 1 machine work on multiple queries as real answers such as tech support wont need such a broad database nor is it under time constraints. | ||
|
Jonoman92
United States9109 Posts
On February 16 2011 14:59 semantics wrote: so much so that it will probably be a long while before we see this on a consumer level. Well not that I know much on the topic, but since this sort of linguistic intelligent technology is in its infancy I don't think it's so unbelievable that in 10 year they'll have something much better than Wilson is in its current state, that also is much more efficiently. Plus consumer processing will continue to improve. Plus, if you have to wait 5 minute for an answer it would still have its uses. Though people may be used to .001 second queries. | ||
|
aztrorisk
United States896 Posts
Ken > Watson First of all, Watson probably hacked the jeopardy system. Not that hard to program a computer to access the jeopardy system if it is connected directly to it. This is why it selected the daily double questions out of the blue. Comeon, it picked something for 800 and the rest for 200. Then it randomly picks something for 600 and then normally picks again when all the daily double are gone. Secondly, Watson had more time to process the question than the human contestants. It receives the question immediately as it appears while both Ken and Brad has to slowly absorb the question until the host is done reading it. This gives Watson 5 extra seconds to process the question while the human contestants has a split second to process it and buzz in. I noticed that Ken and the other guy tried to buzz in sooo many time just a split second too late. Ken > Watson | ||
|
Shinshady
Canada1237 Posts
On February 16 2011 16:01 aztrorisk wrote: Watson is a major disappointment. Ken > Watson First of all, Watson probably hacked the jeopardy system. Not that hard to program a computer to access the jeopardy system if it is connected directly to it. This is why it selected the daily double questions out of the blue. Comeon, it picked something for 800 and the rest for 200. Then it randomly picks something for 600 and then normally picks again when all the daily double are gone. Secondly, Watson had more time to process the question than the human contestants. It receives the question immediately as it appears while both Ken and Brad has to slowly absorb the question until the host is done reading it. This gives Watson 5 extra seconds to process the question while the human contestants has a split second to process it and buzz in. I noticed that Ken and the other guy tried to buzz in sooo many time just a split second too late. Ken > Watson Its not hacking the system... I'm just contemplating this, but I believe through computer learning, that it picked up on the fact or it was coded into Watson, that the daily doubles always appear within that 1200-2000 range, and after it sought them out, it went back to the same style it was playing in the first round. Also human contestants see the question at the same time, but possibly its takes awhile to process it, or Watson has godlike "clicking" speeds, so that the other contestants couldn't buzz in as the host has to finish the question before they can buzz it, whereas if what you say is true, what doesn't stop Watson from buzzing in before the host finishes the question, as he has no voice recognition IIRC. Therefore he wouldnt technically know when the host finishes the question. | ||
Womwomwom
5930 Posts
| ||
|
lixlix
United States482 Posts
As soon as it gets a sizable lead, it switches strategy to picking low value questions to prevent its opponents from coming back and not gambling on DD questions. | ||
|
Zona
40426 Posts
So everyone gets the story they want - IBM gets the "wow the computer is pretty competitive" while the audience/general public still gets the "phew humans are still superior (for now)" But watched part of the episode - must be kind of spooky for the human contestants to play against that thing...just a screen in place of a person. | ||
|
aztrorisk
United States896 Posts
whereas if what you say is true, what doesn't stop Watson from buzzing in before the host finishes the question, as he has no voice recognition IIRC. Last edit: 2011-02-16 16:15:49 Therefore he wouldn't technically know when the host finishes the question. How would watson know when to buzz in even without my hypothesis? Therefore, there MUST be a way that he is notified when the question finished and if he as a certainty rate of over 50%, he will immediately buzz in. I don't know how he knows when the host finishes the question, but all I know is there is some method in which he knows when the host finishes the question and he is able to buzz in immediately if he is done processing and has a certainty rate of higher than 50%. | ||
|
igotmyown
United States4291 Posts
| ||
|
BlackJack
United States10574 Posts
On February 16 2011 19:39 igotmyown wrote: For matches against the machine, if a contestent buzzes in within some small time period (like .25 seconds) and the computer does as well, someone should be randomly selected to answer. Yeah these matches are kind of stupid. They didn't prove they could build a computer that can beat a human at jeopardy, all they've proven is that they've built a robot that can click a buzzer faster than a human.. | ||
|
pirsq
Australia145 Posts
| ||
|
Mig
United States4714 Posts
If Watson gets to choose a category and clue, its first priority is finding any remaining of the three Daily Doubles in a game. These clues allow a contestant to wager a specific dollar amount on the clue without worry of the other two contestants buzzing in. Jennings, Rutter and Watson have a high chance to answer these correctly, so Daily Doubles provide three opportunities for a critical score boost. The Watson Research team studied the historical distribution of Daily Doubles and found they appear most-frequently in the three bottom rows, with the fourth being the most common. Daily Doubles also most frequently appear in the first column. Watson also makes use of even more statistics to dynamically predict their location based on what has been exposed so far in a game. Once the Daily Doubles are off the board, Watson looks for the lowest clue value in a category, for which there are still a significant number of high value clues. Lower value clues help it get the gist of a category with less risk, so that it has a better shot at the high value clues to come. | ||
|
W2
United States1177 Posts
| ||
| ||
