Weak AI beat Chess, now on to Jeopoardy!

[ElectricFetus]

No I stated why I know it can't be done, I never mention this on Sci-forums because I frankly can't stand the subject anymore but I am a Qualified Electrical Engineer aswell as a PT. I did a 3 year course On Electrical Engineering which consisted of split 50/50 Practical and Theory classes.

And that is proof of what? Are you saying I should believe you mealy because your an expert? You know medical experts hired by the tobacco companies said smoking was not harmful, should I believe them? Certainly I would want to examine and verify their expert research instead of taking their expert advise without question. So I would like you to state your expert reasoning, evidence and logic, not simply your opinion alone.

After playing around with mother-boards, power sources, circuits, Soldering-Irons and components for 3 years It was blatantly obvious to me that we are not just systems of transmitters and revievers linked with Wiring.

Well if motherboards and power supply is all you were messing with I don't see how you could not come up with such reasoning! Now if you had spent decades trying the build a thinking conscious machine and fail and derive why it failed and showed with evidence that its failure was because of components to the human mind which can't be replicated, then I would believe you.

You think of humans on a purely physical matter level, you haven't even begun to consider the Quntum Mechanics of the brain.

Quantum mechanics is still physical, and can be replicated, certainly say with quantum computers. Also it unlikely the human brain takes viable advantage of quantum mechanic properties like entanglement or superpositions, which would be very calculation intensive to simulate in a digital Turing machine. Certainly having the brain represented virtually in the Turing machine would require many fold more computing power in the machine then the brain it self produces even assuming a simplistic modem of neural behavior. Perhaps an analogy field programmable gate matrix would be a better system to emulate human intelligence with, it would at least undergo fuzzy logic and neural networking like the brain does in hardware instead of in software.

I seriously hate to burst your Sci-Fi related cyborg desire bubble to have a Robot friend like Optimus -Prime but it really isnt going to happen, you are more than a mere set of circuits.

When have I stated I want Optimus-Prime as a friend? I merely believe the question is open and that we don't know yet if we are more then our circuits, you believe so as a matter of faith, I do not, I mere believe in waiting for more evidence.

Being self aware isnt all human minds are either, being able to comprehend, visualize and actualy appreciate the sensations we recieve is beyond machinery.

How so, explain why.

The Ghost in the Machine is what your Ego and self is.

Again where is this ghost in the machine, is if physical, does it exist in our reality alone, if not were does it exist, where is this other reality? And if does exist in this reality alone then why can't it be replicated simulated or emulated?

 

[spidergoat]

Never say never. The problem with this machine is it has no life experience and it doesn't build a model of the universe it experiences. Neural net architecture combined with an artificial human body with senses and very fast processing speed could attain all the attributes of consciousness.

But we first have to understand how the brain works. When this is accomplished, it will be the final nail in the coffin of theistic religion.

 

[cosmictraveler]

But we first have to understand how the brain works. When this is accomplished, it will be the final nail in the coffin of theistic religion.

I'd disagree because there always will be those, as there are now, that won't ever give in to common sense and the scientifc methold no matter how good machines become in the future.

 

[spidergoat]

True, but the religious explanation for consciousness and the soul will be nullified if we can make our own.

 

[ElectricFetus]

Certainly it could degrade religion if Strong AI are made, but many will refuse to believe these machines are truly aware and conscious, although some may give in but what will be left of the followers will likely become even more fundamentalist and conservative in their thinking. Certainly if the singularity comes humanity will be fractured with those joining the machines and those refusing, even violently so. The machines and transhumans will likely take the high ground (literally) and leave the talking monkeys on this dirty mud-ball of a planet to our own devices.

 

[Captain Kremmen]

That Watson is a Geek of the first order.
Let the machines go to their new planet if they wish.
All they will do is play games of Jeopardy.

 

[spidergoat]

That's like complaining that the Wright brother's first airplane didn't have complimentary beverage service or an in-flight movie.

 

[Sarkus]

True, but the religious explanation for consciousness and the soul will be nullified if we can make our own.

But there are those who say that this artificial consciousness was created by consciousness, and that it is still not possible for consciousness to arise out of "base materials"... so there will still be room for Him upstairs in their thinking.
Plus they could probably seek means and ways to claim that the machine is not truly conscious - although how they could seek to do that I will leave to them when the time comes.

 

[ElectricFetus]

That Watson is a Geek of the first order.
Let the machines go to their new planet if they wish.
All they will do is play games of Jeopardy.

that clearly false as machines do far more today then just play jeopardy, jeopardy is merely something machines had not done before, and it quite clear machines will be able to do even more things in the future.

 

[Captain Kremmen]

Too general.
For example?

 

[ElectricFetus]

Too general.
For example?

Well for example machines connect in a vast networked storing incredible amounts of information which can be access in an instant which you are utilizing right now.

 

[cosmictraveler]

I enjoyed seeing Ken put himself behind the Watson avatar and show only his arms and hands flailing about at the end of the show, that was very funny as was the other guy putting his two fingers up behind the avatar and making the horns symbol, great end to the show.

 

[adoucette]

Too general.
For example?

The amazing part of the computer is that it understood natural language very well even when the questions were phrased in the clever Jeopardy way.

So think of how well it would do if you weren't trying to be clever in your questions?

It's speed and accuracy rate for information was far better than the best Jeopardy players in history and that's impressive when you consider that they were phenomenal at retrieving a wide range of information.

So how would you use it?

Well think how much it could be used for by scientists and researchers to do data mining of scientific, social and business data.

Add in more specific data bases of cutting edge science or business or social issues and then combine that capability with the search engine capability that exists for the Web and with all that data it is a pretty formidable research and exploration tool.

My guess is we will be blown away by how well it is used in the near future and less than a decade from now you will be able to call up on your cell and for a tiny click charge get virtually any question you want answered, instantly and accurately.

Arthur

 

[cosmictraveler]

My guess is we will be blown away by how well it is used in the near future and less than a decade from now you will be able to call up on your cell and for a tiny click charge get virtually any question you want answered, instantly and accurately

You can already do that now if your cell phone is equipped with internet access, which most are for a fee.

 

[ElectricFetus]

Well Watson won, by like 3 times! It certainly knew its stuff and answered many questions that were very tricky, yet there were some questions which would be rudimentary for a human that Watson failed, showing that watson's intelligence (if it can be called that) is austic or grossly deformed from human intelligence: clearly very smart in some ways and very dumb in others, not unexpected for a Turing machine that in no way is trying the simulate human thought.

I noticed some problems with watson designs.
- how could they allow it to get the questions electronically? that not fair, it should have had speak recognition and visual reading, it should have been giving the video output that was sent to the monitors that displayed the question, it would not have taken much computing power the read the words.
- It had no voice recognition so it made no sense for Alex to even talk to it! Voice recognition would have been helpful in hearing the answer the opponents got wrong and thus correcting its own answers.

You can already do that now if your cell phone is equipped with internet access, which most are for a fee.

An existing search engine is nothing compared to what watson could do. Type "this fluid surrounds the nucleus" into a search engine like Google and all you get are links, nothing even displaying the answer directly. Watson would just tell you the answer (cytoplasma).

Type in "who was the first western to make it to india?" do you get back a name? The search engine does not understand what you mean by "western", that is a language cliche, heck the search engine does not even understand that a question has been made, all it does is try to match the text provided to the text on web pages. Watson or rather the DeepQA program that makes up Watson can actually understand the question, it knows that your looking for a western person would traveled to India first.

Heck I bet when they do put Watson or that DeepQA program that makes up Watson online for people it will answer you questions and then some. It will tell you who that person was, when he landed, where he was from, it will then recommend web pages and books on the man and his exploration.

 

[EmptyForceOfChi]

And that is proof of what? Are you saying I should believe you mealy because your an expert? You know medical experts hired by the tobacco companies said smoking was not harmful, should I believe them?

Then I can give no more in-put that is of use to you, all I can do is give opinions and advice based off of my own experiments and experience, or Cite other peoples Data and opinions, I have given my in-put and reasons for my conclusions.


Peace and Guidance

 

[EmptyForceOfChi]

PS.

I wasn't being sarcastic btw I really have no interest in Tinkering with Electronics anymore I see no progress in artificial complexity, Nature is where you need to look for the most amazing complex systems, all humans do is grasp a meek understanding of something in nature and attempt to reverse engineer it. We will never produce anything as Stunning'ly genius as the Human Brain.

The only technology that has even peaked my interest radar briefly was Bluetooth tech, and I am bored with that now too.


Peace

 

[ElectricFetus]

Then I can give no more in-put that is of use to you, all I can do is give opinions and advice based off of my own experiments and experience, or Cite other peoples Data and opinions, I have given my in-put and reasons for my conclusions.

No you can give use actual results from your experiments that verify your claims! You have given us no reason for your conclusions other then that is what you believe and you won't be persuaded otherwise, and such articles of faith is not reasoning.

 

[Sarkus]

... yet there were some questions which would be rudimentary for a human that Watson failed, showing that watson's intelligence (if it can be called that) is austic or grossly deformed from human intelligence: clearly very smart in some ways and very dumb in others, not unexpected for a Turing machine that in no way is trying the simulate human thought.

You need to bear in mind what Watson was attempting to do - and that was understand and interpret complex natural language.
If it doesn't quite understand the nuance of a question it is as fallible as a human in getting the question right or not.
One thing to note, I understand, is that Watson could not refuse to answer a question... it could be beaten to an answer, but if asked it would give one... abstention was not an option for it.
Plus it did not have the answer to everything in its databanks... just most things.

I noticed some problems with watson designs.
- how could they allow it to get the questions electronically? that not fair, it should have had speak recognition and visual reading, it should have been giving the video output that was sent to the monitors that displayed the question, it would not have taken much computing power the read the words.
- It had no voice recognition so it made no sense for Alex to even talk to it! Voice recognition would have been helpful in hearing the answer the opponents got wrong and thus correcting its own answers.

To be honest you're being picky and also misunderstanding the nature of the tech-demonstration.
Speech recognition is entirely separate to what Watson was designed for. Even a basic recognition system can be taught to recognise speech patterns through training, but Watson was not aiming for this tech, but for understanding the language used.
Yes, eventually they will give it a speech recognition front-end, but that is for a later date.
And Alex talking to it was to give the questions to the other participants and also "for show".

An existing search engine is nothing compared to what watson could do. Type "this fluid surrounds the nucleus" into a search engine like Google and all you get are links, nothing even displaying the answer directly. Watson would just tell you the answer (cytoplasma).

Actually, if you read the first few entries of Google for this question (no quotes) you will see the word cytoplasm... at least I did.
The difference is that Google is merely looking for articles with the words in the search bar... Watson is understanding the question and searching for an ANSWER.
I.e. that "cytoplasm" can be seen in the google returns is not recognised by the Google search engine.

But basically Watson's design had few of the problems that you state... the problem lies in your understanding of what Watson was aiming to achieve and/or demonstrate.

 

[Captain Kremmen]

Type in "who was the first western to make it to india?" do you get back a name? The search engine does not understand what you mean by "western", that is a language cliche, heck the search engine does not even understand that a question has been made, all it does is try to match the text provided to the text on web pages. Watson or rather the DeepQA program that makes up Watson can actually understand the question, it knows that your looking for a western person would traveled to India first.

We could let Shadow ask the questions.