[Freddy]

Came across this short story today.  Looks like IBM have worked out how to deal with difficult input (for the computer) and generate an answer.  Although they also say that IBM have not been able to get the machine to reason.

...IBM has issued a new challenge — to the TV game show “Jeopardy!” The company’s latest supercomputer, named Watson, is designed to answer questions well beyond what an Internet search engine can do.

Full story here...

[victorshulist]

Interesting, but still no cigar.    I don't think it has full contextual understanding.   Still just a word indexer, statistical analyzer, word frequency, etc.

[Bragi]

Ye, I think so to. Probably cool though.

[lordjakian]

I don't think it has full contextual understanding.   Still just a word indexer, statistical analyzer, word frequency, etc.

geez victor, how much more do you need to say a program has "full" contextual understanding?

[GT40]

Afraid I missed something. That thing is what we call a computer, with a program inside, right?

Elementary my dear Watson. 

[mendicott]

FYI, I've got an "ibmwatsonbot" at http://twitter.com/ibmwatsonbot monitoring "social buzz" about #IBMWatson ..

You can find Stephen Baker at http://twitter.com/SBFinalJeopardy , author of the upcoming "updateable e-book" about #IBMWatson , "Final Jeopardy: Man vs. Machine and the Quest to Know Everything" http://tinyurl.com/2vmsvvu ..

There is a good article about #IBMWatson , "Building Watson: An Overview of the DeepQA Project", AI Magazine, Vol. 31, No. 3. (2010), by D. Ferrucci, E. Brown, J. Chu-Carroll, et al.

 - Marcus Endicott
   http://twitter.com/mendicott

[Freddy]

Thanks for the links Marcus and welcome to the site 

[mendicott]

Just a quick follow-up to point AIDreamers to my latest blog post attempting to
reverse engineer IBM Watson ..

Marcus L Endicott: How Many PlayStations Make A Watson?
http://www.mendicott.com/2011/01/how-many-playstations-make-watson.html

[Art]

Thanks! That was a really interesting article. I think Watson will handle things wonderfully!

Good stuff!!

[Data]

[DaveMorton]

I guess I'll have to make sure that all of the remotes are either in my direct control, or well hidden, come Monday evening. I'm NOT gonna miss this!

[photostill]

It's amazing today what is done with computers, especially when you look back at where the original started at.

Computers and programming still have a long way to go to successfully imitate the human brain. What is trying to be mimicked is far more complex and until computers make at least a couple of generational jumps, won't be possible.

When you consider that if all the computers made up till the 00's were on line today, they still would not equal the complexity of the brain, getting just one program or one computer to mimic parts of the human personality and thinking capacity is truly amazing.

In March of '90, a maker of microchips did mimic part the brain's complexity with 200,000 neurons linked up by 50 million synaptic connections. That still falls far short of the needed total. They had planned on thousands of chips to create a circuit with a billion neurons and 10^13 synapses (about a tenth of the complexity of the human brain).

Understand, beyond doing a few scripts, I am not a programmer. Still that these can be done, means that some limited areas are advancing but have huge strides yet to go. We may not see what is sought within our lifetimes, even given the speed at which knowledge doubles today.

[DaveMorton]

I'm an active member and moderator over at chatbots.org, and there are currently several members there who are working on, and discussing, their own projects, several of which are along the same lines that were mentioned in that Nova video about Watson. Between NLP processing, and whether it's more important to create "machine knowledge" or to "fully understand the input", some of the discussions get quite intense. As you can guess, the forums there are more technically oriented, and thus, less... Um... I don't want to say "friendly", because the folks there are very friendly, within the scope of the forums. But the guidelines there are much less "relaxed" than here. I love both places. I learn more there, but I have more fun here.

[Freddy]

Hmm I dunno, I think some of Bragi's converstions with other programmers really baked my noodle some times, they are very technical to me.   

There is some technical chat here though, perhaps more in the past.  We've had a lot of philosophical debates too which can really be a lot of fun.  Fresh blood (yikes) is always needed and a good dusting down of our ideas by new people.

Actually though, chatbots org does seem good for programming talk.  In the past places like the VHumans forum had a lot of that kind of thing too, but for some reason that forum doesn't get much action these days, though there is an encouraging trickle starting again recently I found.

It's all worth talking about

[DaveMorton]

Early on in my search for AI and chatbot forums, I checked out the vHumans forum, and found it to be, not only too technical for my (at the time) level of expertise, but also found their replies to my posts "from a noob" to be faintly condescending, and somewhat stand-offish, so I stopped visiting. I don't mind if the topic of discussion is over my head, as long as I have some sort of "frame of reference" to look into and study, so that I can perhaps gain an understanding of what the discussion entails. But when folks are talking about the necessity of having a dozen audio channels, split by frequency - per side(only for stereo, of course. Even more would be "optimal"), for a neural net to be even considered to be viable, it's time to look elsewhere for pointers for my AIML chatbot. I'm a pretty smart individual, but I'm not quite at the doctorate level, when it comes to AI.