17 replies to this topic

[Elus]

http://www.youtube.com/watch?v=FC3IryWr4c8


http://www.internetb...computer/56872/

Pretty incredible.

[albedo]

"On Wednesday, IBM, along with Nuance Communications Inc. and the Columbia University and University of Maryland medical schools, announced that they are developing Watson as a diagnostic tool that can help doctors identify diseases and recommend treatments."

http://www.time.com/...2049826,00.html

[Marios Kyriazis]

I still not think that this achievement is anywhere near mimicking human abilities or human mind. Here are some selected extracts from Ben Goertzel's comments:



Watson: Supercharged Search Engine or Prototype Robot Overlord?
By: Ben Goertzel
H+ magazine, February 17, 2011
full article at: http://hplusmagazine...robot-overlord/

My initial reaction to reading about IBM's "Watson" supercomputer and software was a big fat ho-hum. "OK," I figured, "a program that plays Jeopardy! may be impressive to Joe Blow in the street, but I'm an AI guru so I know pretty much exactly what kind of specialized trickery they're using under the hood. It's not really a high-level mind, just a fancy database lookup system."......

In the final analysis, Watson didn't seem human at all - its IBM overlords didn't think to program it to sound excited or to celebrate its victory. While the audience cheered Watson, the champ itself remained impassive, precisely as befitting a specialized question-answering system without any emotion module.....

But who is this impassive champion, really? A mere supercharged search engine, or a prototype robot overlord?

A lot closer to the former, for sure. Watson 2.0, if there is one, may make fewer dumb mistakes - but it's not going to march out of the Jeopardy! TV studio and start taking over human jobs, winning Nobel Prizes, building femtofactories and spawning Singularities. [...]

[Elus]

Well, Goertzel also says, "But even so, it was pretty bloody cool to see it up there on stage, defeating humans in a battle of wits created purely by humans for humans — playing by the human rules and winning."

We shouldn't overhype the achievement, but let's not downplay it either. This isn't something that happens every day.

Edited by Elus, 22 February 2011 - 01:38 PM.

[niner]

Well, Goertzel also says, "But even so, it was pretty bloody cool to see it up there on stage, defeating humans in a battle of wits created purely by humans for humans — playing by the human rules and winning."

We shouldn't overhype the achievement, but let's now downplay it either. This isn't something that happens every day.

No, it isn't. I think it was a great achievement by humanity, building a machine that can outplay human Jeopardy! experts. Watson is certainly not a human-level general intelligence, and no one is claiming that it is. It is, however, an amazing collection of hardware and software. What we should be doing is projecting this device about ten years into the future. Instead of a room full of hardware, how large (and expensive) a computer will it need? Certainly smaller and cheaper. Handheld? I couldn't say. With a full decade of software development implementing the latest in knowledge representation, where might it go? Better and smarter is all I can say. Superhuman? Well, it's already superhuman in the Jeopardy! domain. Expect to see a widening of its domain. I look forward to the day when the successors of Watson read the entirety of the scientific literature, start recommending what experiments to do next, and making the connections that we are presently missing. I think a scenario like that is inevitable, and I expect to see it inside of 20 years.

[the thing]

It is pretty impressive, but it seemed like most of it's success came from it's ability to buzz in much faster than humans (Jennings especially seemed little frustrated).

[Marios Kyriazis]

I look forward to the day when the successors of Watson read the entirety of the scientific literature, start recommending what experiments to do next...

Fully agree and hope. But this means working with humans (i.e. to enhance us), not to replace us.

I think a scenario like that is inevitable, and I expect to see it inside of 20 years.

...or less.

[Elus]

Also, as Valkyrie_Ice mentioned earlier (I don't know where, perhaps some other forum post), we could could organize medical knowledge in a central database in such a way that Watson (and his descendants) are able to read and process it more quickly. This will help increase Watson's utility for things such as medical diagnosis.

[albedo]

Yes Elus, the announcement IBM made and which I referred to in my previous post goes along these lines. Also there are other important industries this technology can be applied to help humans. As with chess, I found this new achievement impressive.

[VidX]

As wrote in some other thread, we became very "spoiled" in regards to technologies.. This is VERY impressive however you look at it.

[albedo]

I agree VidX and not only for being spoiled. I find that today, being completely overwhelmed by technology information and all sort of gadgets, it it sometime hard to sort the signal out of the noise and arrive to grasp signification of small steps generating real progress. What they did is a small step, true, but relevant for the industry and the human-computer interface field.

[VidX]

Kurzweil on Watson ( few mins into the interview):

www.youtube.com/watch?v=J8_jz9QlrUI&feature=relmfu[


p.s. - ok, one more time - how the hell is it possible to embedd youtube?? Used to be video...url...video.. doesn't work.

Edited by VidX, 10 March 2011 - 06:01 PM.

[Elus]

p.s. - ok, one more time - how the hell is it possible to embedd youtube?? Used to be video...url...video.. doesn't work.

Here:

http://www.youtube.com/watch?v=J8_jz9QlrUI



This is the code you need to use: [.media]Youtube link address here[./media] (minus the periods in the media tags).

Edited by Elus, 11 March 2011 - 07:24 AM.

[chris w]

They will be putting Watson to use analysing medical data

IBM didn’t develop a supercomputer capable of winning at Jeopardy just to pull in game show money. The company announced a deal this morning with health insurance giant WellPoint to use the Watson supercomputer to parse health care information.

It’s the first commercial application for Watson, the computer most famous for taking on Jeopardy champions and whipping them at their own game.

IBM and WellPoint have agreed to use Watson to analyze healthcare data for the insurance company’s customers to help doctors and nurses diagnose and treat conditions patients could have.

"There are breathtaking advances in medical science and clinical knowledge, however; this clinical information is not always used in the care of patients. Imagine having the ability to take in all the information around a patient's medical care -- symptoms, findings, patient interviews and diagnostic studies. Then, imagine using Watson analytic capabilities to consider all of the prior cases, the state-of-the-art clinical knowledge in the medical literature and clinical best practices to help a physician advance a diagnosis and guide a course of treatment," said Dr. Sam Nussbaum, WellPoint's Chief Medical Officer, in a release. "We believe this will be an invaluable resource for our partnering physicians and will dramatically enhance the quality and effectiveness of medical care they deliver to our members."

The companies did not say how much WellPoint would pay for access to Watson’s computing power. But it’s a fair bet the largest publicly-traded health insurer is forking over more than the $1 million the computer won on Jeopardy.

http://www.portfolio...k-for-wellpoint


I don't know why only this much shows up, I can't edit it for some reason, enter the link for the rest of the text

Edited by chris w, 12 September 2011 - 10:11 PM.

[chris w]

DP

Edited by chris w, 12 September 2011 - 10:10 PM.

[albedo]

WellPoint and IBM Announce Agreement to Put Watson to Work in Health Care

http://www-03.ibm.co...lease/35402.wss

[albedo]

I found Ray Kurzweil's reply to Paul Allend criticism interesting, also showing what IBM has been achieving with Watson and the work with Nuance (a company founded by Ray) to have Watson read huge amounts of medical articles to support medical diagnostics. Only this, independently from (important, though) philosophical discussions about passing or not the Turing test, is simply extraordinary.

[steampoweredgod]

It is pretty impressive, but it seemed like most of it's success came from it's ability to buzz in much faster than humans (Jennings especially seemed little frustrated).

IIRC, I read somewhere that it was actually not able to beat human's peak buzz speed reaction, and it won despite that.