IBM Computer Program To Take On 'Jeopardy!' 213
longacre writes "I.B.M. plans to announce Monday that it is in the final stages of completing a computer program to compete against human 'Jeopardy!' contestants. If the program beats the humans, the field of artificial intelligence will have made a leap forward. ... The team is aiming not at a true thinking machine but at a new class of software that can 'understand' human questions and respond to them correctly. Such a program would have enormous economic implications. ... The proposed contest is an effort by I.B.M. to prove that its researchers can make significant technical progress by picking "grand challenges" like its early chess foray. The new bid is based on three years of work by a team that has grown to 20 experts in fields like natural language processing, machine learning and information retrieval. ... Under the rules of the match that the company has negotiated with the 'Jeopardy!' producers, the computer will not have to emulate all human qualities. It will receive questions as electronic text. The human contestants will both see the text of each question and hear it spoken by the show's host, Alex Trebek. ... Mr. Friedman added that they were also thinking about whom the human contestants should be and were considering inviting Ken Jennings, the 'Jeopardy!' contestant who won 74 consecutive times and collected $2.52 million in 2004."
Logs (Score:3, Interesting)
logs or it didn't happen! what they did to Kasparov was bullshit! Seriously if this magically gets better at 1/2 time, the least they can do is show the logs
Buzzing In (Score:3, Interesting)
Re:Leap Forward? (Score:3, Interesting)
What msbmsb said. Also, this sounds similar to the problem already tackled (and aced) by Google Sets (or whatever they call it now). That's the feature where you give it some members of a set you have in mind (but you don't tell it what it's a set *of*) and it outputs more members from that set. For example, you give it "apple, orange, banana", and it gives you "grape, strawberry, lime". I'm guessing the way Google sets works is:
1) Run a search on all input phrases.
2) Find the most common statistically-improbable phrase in all of the results.
3) Run a search on the phrase derived in 2.
4) Output the most common statisticall-improbable phrases found in three which were not given as inputs.
In the example you gave, there will be more search results with "hamlet", "playwright", and "Shakespeare" than "hamlet", "playwright", and some other SIP.
There are, of course, many other cases where it will get tripped up (like puns and rhyming), but that's not one of them.
Re:I'd take Jennings (Score:3, Interesting)
The problem they might run into is the speed of pressing the button to respond. I would imagine the computer would be able to beat the human every time it knew the answer.
This is actually where I think the humans have an advantage. They can press the button just because they think that they WILL know the answer in the time allotted.
Watson may be designed to predict its own ability to answer. But to allow it to just press the button, then use the entire time limit to find it would not be fair...
Total geek-gasm (Score:3, Interesting)
The way to deal with such problems, Dr. Ferrucci said, is to improve the programâ(TM)s ability to understand the way Jeopardy! clues are offered. The complexity of the challenge is underscored by the subtlety involved in capturing the exact meaning of a spoken sentence. For example, the sentence "I never said she stole my money" can have seven different meanings depending on which word is stressed. "We love those sentences," Dr. Nyberg said. "Those are the ones we talk about when weâ(TM)re sitting around having beers after work."
Seriously guys, I just had a geek-gasm. Anyone else?
State of the Art Voice Recognition (Score:5, Interesting)
We were just talking about this in another thread... A lot o the comments here have been that natural language software isn't that great.
This isn't at all true. Today, understanding verbal and written communication is done by state of the art computers and programs at a rate about equal to human listeners and readers. Where a computer doesn't particularly excel is in parsing that language, mostly because a computer doesn't have access to our culture in order to absorb context, but context can still be added.
Here's an example from Ray Kurzweil's book "Age of Spiritual Machines." He talks about a phrase famously given to a language parsing program that goes thus: "Times flies like an arrow." This phrase can be understood in various ways:
"* The common simile: time moves quickly just like an arrow does;
* measure the speed of flies like you would measure that of an arrow (thus interpreted as an imperative) - i.e. (You should) time flies as you would (time) an arrow;
* measure the speed of flies like an arrow would - i.e. Time flies in the same way that an arrow would (time them);
* measure the speed of flies that are like arrows - i.e. Time those flies that are like arrows;
* all of a type of flying insect, "time-flies," collectively enjoys a single arrow (compare Fruit flies like a banana);
* each of a type of flying insect, "time-flies," individually enjoys a different arrow (similar comparison applies);
* A concrete object, for example the magazine, Time, travels through the air in an arrow-like manner."
(from: http://en.wikipedia.org/wiki/Natural_language_processing [wikipedia.org])
With a few facts it becomes obvious which is correct and which isn't. Tell the computer that there's no such thing as a 'time fly' and that flies don't time things, and that flies aren't sophisticated enough to like things in an affectionate manner, etc., and the correct interpretation soon becomes clear.
So, if you think a computer can't understand both written and verbal communication and then parse it quickly enough to answer the questions I will have to strenuously disagree. These challenges are quickly being overcome on the bleeding edge of the art. But since this perception persists that the state of the art is somehow bad, because Joe down the street messed with some free-ware language software that worked poorly -- I think a lot of people are in for a surprise, and winning Jeopardy in this manner is really the perfect way to show it off. Can't wait to see the Youtube clips.