IBM's "Watson" finally ready for prime-time Jeopardy • Hacker News Board

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#41

Post by adsr » Thu, Feb 10, 2011, 5:07 PM UTC

JoeAltmaier wrote: post #4

cryptoz wrote: post #2

Humanity is closing in on building machines to pass the Turing Test. Watson beat both champs in the demo round a few weeks ago and I fully expect him - yes, him - to beat them in the match coming up. I can't wait to watch this live. What an epic moment in the process of moving slowly from weak AI -> strong AI. Watson can do amazing NLP (and presumably Machine Learning), which is something that the general public perc…

Watson is a simulation, a clever puppet designed to mimic human behavior. We have no more to fear from it, than from a punch and judy show. Cleverer and cleverer puppets will not become intelligent machines. A dedicated effort to build a massive neural cluster simulation might become intelligent, but it will think such incomprehensible thoughts as to be fairly useless to us (think intelligent rhododendon; what would…

I agree about the simulation part, does it have a sense of self? Or can pose questions to itself such as "I think therefore I am"?

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#42

Post by derwiki » Thu, Feb 10, 2011, 5:08 PM UTC

I think it would be cool to see IBM's creation versus a creation from Google versus (a person or another machine), although I doubt another company would want to make such a risky move. If Watson wins, huge PR win; if Watson loses, still a pretty big PR win.

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#43

Post by kenjackson » Thu, Feb 10, 2011, 5:10 PM UTC

kls wrote: post #36

kenjackson wrote: post #22

How much of the Watson code is Java?

I don't know the percentages but Watson is built on UIAM http://uima.apache.org/ which has bits in C++ and bits in Java. From there I am sure different teams used language bindings to use the language that best suited their particular needs to accomplish their tasks. Not that I know it to be true, but I would not be at all surprised to find R, Haskel and even some lisp in their somewhere.

Thanks for the link. That looks really impressive. For IBM this is presumably PR for their products/services. I do hope they disclose more of the architecture as I'd love to build out a system like this, even on a smaller scale (takes 6 hours to come back with an answer rather than three seconds).

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#44

Post by bgurupra » Thu, Feb 10, 2011, 5:11 PM UTC

glhaynes wrote: post #40

bgurupra wrote: post #33

http://www-943.ibm.com/innovation/us/watson/

They can win at Jeopardy but they can't build a URL structure without annoying machine/network identifiers. :)

that is actually my bad , this URL works too http://www.ibm.com/innovation/us/watson/ - basically this sends a request to the main front end proxy and because IBM hosts like a billion URLs the front end proxy redirects to the appropriate cluster

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#45

Post by kenjackson » Thu, Feb 10, 2011, 5:14 PM UTC

gcheong wrote: post #34

I've looked around a bit but couldn't find any information on how the game state and clues are input to Watson. Are they typed in on the fly? Pre-loaded and just revealed as clues are selected? Is it parsing speech and reading the board visually?

I've heard (from Engadget) that it is given text on the fly. It does not do speech or visual recognition. But it gets the question the same time the contestants see them.

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#46

Post by kls » Thu, Feb 10, 2011, 5:15 PM UTC

arethuza wrote: post #38

kls wrote: post #25

Earlier quoted context omitted.

No disrespect intended but, I think you are lacking a fundamental understanding of machine learning, genetic algorithms, neural networks and AI in general. I was introduced to Watson when I took a sub-project from IBM related to Watson, I had no idea what it was untill taking the contract, at which point I was introduced to Watson and it is an impressive feat, it is with out a doubt the state of the art in NLP AI. Su…

I think anyone who has worked in the field of AI would admit there are a lot of philosophical questions around the definition of intelligence and it's relationship to algorithmic problem solving. One of my favourite thought experiments in the area being Searle's Chinese Room: http://en.wikipedia.org/wiki/Chinese_room

Sure, intelligence is much like religion or faith, what it is is a very personal matter as well a philosophical matter to each observer. To me if a machine can make a logical decision that it arrives at via it's own conclusion independent of human involvement (after the initial construction). It can be said to be intelligent. I personally think we have already achieved intelligent machines. Granted not intelligent in the traditional biological sense but intelligent in their own right. I do believe that we are close to hyper-intelligent machines than many think we are. I personally think it will be the next big step. I also believe that it will fundamentally and irrevocably change humanity. It is a Pandora's box that we will not know the ramifications of until we pass that event horizon, it is much like Schrodinger cat, full of possibilities but until we experience the event we will not know how it affects humanity.

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#47

Post by rst » Thu, Feb 10, 2011, 5:17 PM UTC

cryptoz wrote: post #2

Humanity is closing in on building machines to pass the Turing Test. Watson beat both champs in the demo round a few weeks ago and I fully expect him - yes, him - to beat them in the match coming up. I can't wait to watch this live. What an epic moment in the process of moving slowly from weak AI -> strong AI. Watson can do amazing NLP (and presumably Machine Learning), which is something that the general public perc…

To seem to me like a person, Watson would need more than intelligence, it would need, well... personality.

Watson will answer questions put to it for as long as the power is on. But it will never do anything else. It has no other desires, no motives, no interests. Its planning skills are limited to game strategy. Its conversational skills are limited to preprogrammed social niceties which probably don't extend beyond the context of the game.

You can make a decent case that it's intelligent --- but it's an intelligent machine. You'd have to add all of the above, and probably more, in order to get something that you could interact with even on-line as a person.

[EDIT: I don't mean to deny the possibility of a machine that I would feel comfortable calling a person --- but I saw the Watson demo run, and I haven't seen one yet.]

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#48

Post by mhb » Thu, Feb 10, 2011, 5:20 PM UTC

Maybe they should give it the answer "The specifications for a machine to beat you at Jeopardy."

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#49

Post by kls » Thu, Feb 10, 2011, 5:22 PM UTC

kenjackson wrote: post #43

kls wrote: post #36

Earlier quoted context omitted.

I don't know the percentages but Watson is built on UIAM http://uima.apache.org/ which has bits in C++ and bits in Java. From there I am sure different teams used language bindings to use the language that best suited their particular needs to accomplish their tasks. Not that I know it to be true, but I would not be at all surprised to find R, Haskel and even some lisp in their somewhere.

Thanks for the link. That looks really impressive. For IBM this is presumably PR for their products/services. I do hope they disclose more of the architecture as I'd love to build out a system like this, even on a smaller scale (takes 6 hours to come back with an answer rather than three seconds).

My understanding that this is the beginning of their DeepQA product line, my assumption is that the research and portions of the technology will be used for business insight and analytical to answer what if type questions.

,Just to be clear I don't work for IBM, and I do not know their intentions for the project, but I do take projects from IBM and have taken projects related to Watson, but I do not know their plans to monetize Watson, the former is just pure speculation on my part.

Re: IBM's "Watson" finally ready for prime-time Jeopardy

#50

Post by hencq » Thu, Feb 10, 2011, 5:25 PM UTC

cryptoz wrote: post #9

JoeAltmaier wrote: post #4

Earlier quoted context omitted.

Watson is a simulation, a clever puppet designed to mimic human behavior. We have no more to fear from it, than from a punch and judy show. Cleverer and cleverer puppets will not become intelligent machines. A dedicated effort to build a massive neural cluster simulation might become intelligent, but it will think such incomprehensible thoughts as to be fairly useless to us (think intelligent rhododendon; what would…

Watson is not a simulation any more than your brain is a simulation. Your brain is ( likely ) mostly just algorithms and data structures, albeit on an enormous scale. This could be wrong, but I'm betting not. What else could the brain possibly be? It's not magic. > We will know when Watson is dangerous when it feels fear, angst, want. Not just an algorithm to sort facts and simulate speech. Nah. Look at good science…

Well HAL 9000, while cold and calculating seems to be scared when he's being turned off. So I'm not sure I agree there are no emotions. Who knows, maybe things like emotion just appear as side effects of the simulation once it becomes complex enough.