Human level AI: One step closer.

[Firehazard159]

This innovation is a huge step in the right direction towards developing the AI we all dream of. Basil, the robot - so awesome.

Start here:

http://www.westword.com/2008-12-18/...t-us-ready-for-basil-the-robot-of-our-dreams/

Then here, for action:

http://www.youtube.com/user/gamma2video

Lastly, here, for their official website:

http://gamma-two.com/bsl.html

 

[kubikub]

heh! Great links, I somehow find Basil strangely cute in a robot-y sort of way.

I also enjoyed this quote from the article.

"We're those evil science types," jokes Jim. "We don't have feelings."

Spoken like a true INTP

 

[Artifice Orisit]

Baby steps, just baby steps.

A couple of fanatics working out of their basement is the cutting edge of technology?
...
I doubt it, after all AI is the new nuke.
Think about it.

Troops are expensive, they need training, equipment, provisions, transport, accommodations, psychological support, relativity humane working conditions, medical care, and when they come back they'll expect to be looked after in their old age.
In the economy of war, artificial troops are a game breaker, they would enable the USA to outnumber China's ground army (the largest in the world) at a ratio of 3/1, easily.
Troops from a factory, spent like ammunition.

Even if robots are more expensive, when a human gets wounded how long is the recovery time? How long for a robot?

 

[Cogwulf]

I'm not a computer scientist and I probably don't have any of the neccesary knowledge to comment upon this, but I think software will never be able to be able to produce AI regardless of the power of the computer. I believe it is impossible to create a program which creates original responses, it is possible to program something to design a respsonse but it can't create outside of what the program allows. I think to create AI we should be trying to create a new form of electronic circuit that replicates an animal brain rather than create software to recreate an animal brain inside a logic circuit

 

[kantor1003]

heh! Great links, I somehow find Basil strangely cute in a robot-y sort of way.

Maybe you are the first human to develop robotophilia

 

[Artifice Orisit]

Maybe you are the first human to develop robotophilia

Lol, the first? You've got to be kidding.

What is a sex toy if not a simple robotic surrogate?

Robots get around

 

[Deleted member 1424]

I'm not a computer scientist and I probably don't have any of the necessary knowledge to comment upon this, but I think software will never be able to be able to produce AI regardless of the power of the computer. I believe it is impossible to create a program which creates original responses, it is possible to program something to design a response but it can't create outside of what the program allows. I think to create AI we should be trying to create a new form of electronic circuit that replicates an animal brain rather than create software to recreate an animal brain inside a logic circuit

*ahem* May I introduce you to the autonomous robot Sprocket. I can personally attest that he is capable of seemingly original responses. He can recognize his surroundings, including scenery and humans. He is capable of making assumptions about his environment and act accordingly. At the very least he is an ingenious little piece of tech and extensive programming. I met him at the X-prize a few years ago and was blown away.

http://www.youtube.com/watch?v=YnzDC0t2Zyo

"I can't do a Christian Bale impression unless you give me a chainsaw and a pair of sneakers." lmao

http://www.youtube.com/watch?v=TmZVnhDZI9U

You may or may not care, but they guy talking is my brother. While we were there we asked sprocket several questions twice, in which he replied similarly, but not exactly as he did before. In fact he recognized that we were trying to test the extent of his autonomy.



http://www.youtube.com/watch?v=MZtxXp3qIOo

A longer video I found on youtube that displays his capabilities a bit better. Also he does move around a lot on his own, it's just not really shown in the video. In fact at the X-prize his 'handlers' pretty much let him do his own thing and wander around. I suppose there is a possibility that Sprocket is remote controlled, but it didn't seem like it to me.


And before you robotphilliacs get any ideas; Sprocket is mine so back off bitches!

 

[Cogwulf]

*ahem* May I introduce you to the autonomous robot Sprocket. I can personally attest that he is capable of seemingly original responses. He can recognize his surroundings, including scenery and humans. He is capable of making assumptions about his environment and act accordingly. At the very least he is an ingenious little piece of tech and extensive programming. I met him at the X-prize a few years ago and was blown away.

It is quite amazing, but the robot is only doing what it has been programmed to do. There's just a big database of information, a speech recognition engine, a grammar engine, and an extensive set of rules to tell it what to do with information. There will also be a set of rules to allow it to learn new information for its database and to expand its grammar engine. It cannot do anything that it has not been specifically programmed to do. I am sure that given enough time I could find questions and statements it could not respond to despite what information it was given.
I think this robot could be adapted to pass the turing test, but only if it was also programmed to behave as if it was not fluent in the language the test was carried out in, and if the test was quite informal.

 

[Agent Intellect]

You should check out the work they're doing with Project Blue Brain.

http://seedmagazine.com/content/article/out_of_the_blue/
http://singularityhub.com/2009/04/3...g-like-the-real-thing-the-blue-brain-project/
http://www.hss.caltech.edu/~steve/markham.pdf (lengthy detailed site)

Also, they can get novel ideas from computers by running genetic algorithms.

http://www.popsci.com/scitech/article/2009-08/evolving-robots-learn-lie-hide-resources-each-other

And even robots that can "think for themselves".

http://www.reuters.com/article/technologyNews/idUSTRE53167K20090403

 

[Firehazard159]

Baby steps, just baby steps.

A couple of fanatics working out of their basement is the cutting edge of technology?
...
I doubt it, after all AI is the new nuke.
Think about it.

Troops are expensive, they need training, equipment, provisions, transport, accommodations, psychological support, relativity humane working conditions, medical care, and when they come back they'll expect to be looked after in their old age.
In the economy of war, artificial troops are a game breaker, they would enable the USA to outnumber China's ground army (the largest in the world) at a ratio of 3/1, easily.
Troops from a factory, spent like ammunition.

Even if robots are more expensive, when a human gets wounded how long is the recovery time? How long for a robot?

Computers, invented in a garage by two guys just tinkering around with an idea they had? Nah.

Oh wait.

Really, you don't think innovations can come from home science? Or even non-government driven science?

Something doesn't add up with that thought, me thinks. So, be skeptical all you want, but until you've got some proof of a greater cutting edge development, I'm thinking this *is* the next step. It's not like the tinkerers are idiots, they're pretty heavily educated.

And Adaire, why is all I can find on sprocket just those youtube videos? Google, wikipedia, all turn up nothing, to research into how he works, which makes me think he's just a remote controlled voice box (well, possibly with autonomous movement, but the speaker seems more likely to be human, especially with all the tonal inflections and what he speaks about). *shrugs*

 

[Deleted member 1424]

It is quite amazing, but the robot is only doing what it has been programmed to do.

Well one could argue that humans only do what their genes and environment have programmed them to do.

Either way, I'm not saying he is an actual strong AI. He's clearly been designed for show purposes, and to look like he has a personality. I do not think he could operate in his own interests, or chose a purpose for himself. I doubt he'd be entertaining humans if he could. He's just an example of how advanced we are technologically. Sure, he's an imitation consciousness, but we are getting closer.

Organic tissue and silicon chips are just [potential] mediums for consciousness. Just as surely as you can have lakes of water and methane, you can have consciousness based on silicon and carbon; provided you have the right conditions.

@fire
http://www.lockheedmartin.com/index.html

This is the company that designed him. He features in a few articles on the website, though your not going to find anything that details his design. Lockheed Martin is a defense contractor, so I doubt they let such details outside the company. I also doubt such a large company would bother with such a hoax.

Computers, invented in a garage by two guys just tinkering around with an idea they had? Nah.

Oh wait.

Really, you don't think innovations can come from home science? Or even non-government driven science?

The further we advance the more difficult is becomes. That's the reason we haven't had any new Teslas, Einsteins or Newtons recently. It takes an incredible amount of specialized knowledge and in many fields to create advancements similar to the early days. You no longer have the singular mad scientist with brilliant ideas. You need a collaboration of incredibly intelligent people (possibly hundreds). Science and technology becomes exponentially more complex as you progress. If we don't start to artificially expand human intelligence we will hit a plateau of sorts where we will not be able to advance.

Advancement nowadays needs an incredible amount of cognitive and fiscal resources, it's not something two guys in a garage can do by themselves.

edit: Thanks for the links AI.

 

[Kuu]

Advancement nowadays needs an incredible amount of cognitive and fiscal resources, it's not something two guys in a garage can do by themselves.

I disagree with you and Cog. There is a difference between technical advancement and conceptual advancement. A lot of technological developments happen only AFTER some (very knowledgeable) dude in his house writes a book with some ground shaking concepts. Theory comes first on many occasions. And it is this conceptual, not technical, advancement that we need in many fields in which we are currently stuck. Just take a look at the history of physics in the 20th century... theoretical predictions always came before the proofs themselves. Witricity (currently in development) is a technology that could have been developed with the knowledge and technology of 60 years ago, if somebody would have thought of it... but they simply didn't.

What kind of INTPs are you! Aren't we the first to complain that we could change the world as it is today, if only people listened to our theories? *is shocked*

I do find this more fascinating than a lot of flashier robots, because I agree that it is possible to create a robot that learns, and that a highly complex system like cognition is an emergent system bound by very elementary rules, which can work irrelevantly of the organic or electronic nature of the system of implementation. It is a conceptual problem which I always thought was completely put aside... quite irritatingly... and I'm glad someone is actually addressing it.

Thanks firehazard.

 

[Deleted member 1424]

I disagree with you and Cog. There is a difference between technical advancement and conceptual advancement. A lot of technological developments happen only AFTER some (very knowledgeable) dude in his house writes a book with some ground shaking concepts.

More likely a company takes a gamble on a preexisting, but untapped concept and funds it. Anyway we're talking about building a robot here. If you think what two people build in a minimal facility can compete with a whole science department with a multimillion dollar budget; then you're just being delusional.

R&D costs money and humans have a limited capacity when working alone without support. Simple as that.

Sure an individual can come up with an interesting concept or idea, but it cannot be perfected and implemented by that singular individual; especially when as you reach higher levels of complexity.

Theory comes first on many occasions. And it is this conceptual, not technical, advancement that we need in many fields in which we are currently stuck.

Like the technical aspects concepts are built upon pre-existing foundations and they grow in complexity as time goes by. The human mind has a limit (for now), and singular genius can no longer shake the world as it once did.

Just take a look at the history of physics in the 20th century... theoretical predictions always came before the proofs themselves. Witricity (currently in development) is a technology that could have been developed with the knowledge and technology of 60 years ago, if somebody would have thought of it... but they simply didn't.

I believe Tesla first discovered this. Nevertheless wireless energy is an old concept, that was just never funded until now.



edit: Tesla fixed

 

[Kuu]

More likely a company takes a gamble on a preexisting, but untapped concept and funds it.

You trust corporations to advance science for science's sake?

Anyway we're talking about building a robot here. If you think what two people build in a minimal facility can compete with a whole science department with a multimillion dollar budget; then you're just being delusional.

R&D costs money and humans have a limited capacity when working alone without support. Simple as that.

Sure an individual can come up with an interesting concept or idea, but it cannot be perfected and implemented by that singular individual; especially when as you reach higher levels of complexity.

AI =/= Robot. I'm also not expecting them to do it all themselves, just setting the conceptual foundations....

Torrents come to mind. Everyone uses them now, but they all started from some guy and his friend in a shabby apartment.... with a novel solution to the same old problem.

I believe Tesla first discovered this (admittedly while looking for a means of wireless communication, not energy). However his research was not funded because of the discovery of radio waves. Nevertheless wireless energy is an old concept, that was just never funded until now.

I wasn't talking about the Tesla kind. (and also, radio waves were discovered before Tesla's Wardenclyffe Project even started.... and the research stopped being funded because JP Morgan wanted to make money of it instead of giving free electricity to the world... )
Its a different, new method. that uses the electromagnetic near field and resonance.


And I don't know where you got the idea that I said it was easy to come up with new things... I said no such thing.

 

[Firehazard159]

It takes an incredible amount of specialized knowledge

Jim has a PhD in Computer Science from the University of Virginia, as well as a BS in Computer Science and Engineering from the University of Colorado at Denver.

He has been actively involved in researching machine based cognition for 20 years, and has focused on embedded intelligence since the late 1990's. His areas of expertise include intelligent systems, robotics, embedded real-time systems, and planning in uncertain domains. He has been awarded several US and European patents for robotic systems, and is the author of numerous technical publications. In addition, he has been the session organizer for several conferences and symposia, and is the co-author (along with Louise Gunderson) of the recently released technical book "Robots, Reasoning, and Reification" on machine cognition and robotics.

Jim is the Cognitive Systems Architect at Gamma Two, Inc.

and in many fields to create advancements similar to the early days.

Louise has broad range of educational accomplishments, including a PhD in Systems Engineering from the University of Virginia, a MS in Environmental Science and a BA in Biology from the University of Colorado at Denver, as well as BA in Chemistry from the University of California at Berkeley.

Her areas of expertise include the automatic extraction of information from complex data streams, geospatial analysis of satellite data, data mining, human preference forecasting, and predictive modeling of human behavior. She is the author of over 20 technical publications, including book chapters and archival journal articles. Along with James Gunderson, she has completed a technical book which was published by Springer-Verlag in late 2008.






Seems like your requirements are met, to me :P

 

[dents]

... There's just a big database of information, a speech recognition engine, a grammar engine, and an extensive set of rules to tell it what to do with information. ... I am sure that given enough time I could find questions and statements it could not respond to despite what information it was given. ...

How is that any different from you? Your brain is just a big database of common sense built over {YOUR_AGE} years of constant interaction with your environment. You have pretty good information processing circuitry (audio, video, temperature, tactile, etc), a language engine and a fairly short list of primary goals (eat, sleep, etc). Everything you do, and will ever do, is in some way related to satisfying those.

Your brain cannot physically be more powerful than a Turing machine, the only difference is you have smaller/better/more hardware. For now.

 

[Artifice Orisit]

NO WAIT! You've got it all wrong.
I'm just saying it's likely there's stuff out there we don't know about.

I'm not saying innovations can't come from garage mad science!
Nor do I think this particular couple are idiots, far from it in fact.

 

[Firehazard159]

NO WAIT! You've got it all wrong.
I'm just saying it's likely there's stuff out there we don't know about.

I'm not saying innovations can't come from garage mad science!
Nor do I think this particular couple are idiots, far from it in fact.

Hehe, I kind of figured you were going there. It seems a bit conspiracy theorist though too, and kind of not at the same time (Obviously, I'm torn.)

However, I'm possibly going to get top secret clearance in the american air force, so maybe I'll get to find out about some of that stuff (Not joking either )

/me hopes the stargate in co. springs is real.

Lulz ^

 

[Cogwulf]

How is that any different from you? Your brain is just a big database of common sense built over {YOUR_AGE} years of constant interaction with your environment. You have pretty good information processing circuitry (audio, video, temperature, tactile, etc), a language engine and a fairly short list of primary goals (eat, sleep, etc). Everything you do, and will ever do, is in some way related to satisfying those.

Your brain cannot physically be more powerful than a Turing machine, the only difference is you have smaller/better/more hardware. For now.

I suppose it all comes down to how you define these things. Is a brain actually anything more than a computer?

 

[Artifice Orisit]

/me hopes the stargate in co. springs is real.

They've got a door that says "Stargate Control Room" ...it's a closet.
Or is it?

(here-say from a doco about the show, I think, memory is fallible)

 

[Firehazard159]

They've got a door that says "Stargate Control Room" ...it's a closet.
Or is it?

(here-say from a doco about the show, I think, memory is fallible)

If you guys don't hear from me after I go through basic, just assume I went off-world.

 

[kantor1003]

Lol, the first? You've got to be kidding.

What is a sex toy if not a simple robotic surrogate?

Robots get around

Hmm, but thats more a poor imitation of a male penis. I think there is a difference between that and our robot friend here as he can't really please a woman sexually, yet, and isn't made for it...so, the women that likes their dildo doesn't have robotophilia, they just use it like a tool to stimulate themselves.

 

[Cassandra]

Amazing!

I really want that book ( Robots, Reasoning and Reification) now...but, I just checked, it costs between 77 and 130 dollars )

Maybe it's at the library.

From what I've read of it:

Spoiler

Robots, Reasoning, and Reification focuses on a critical obstacle that is preventing the development of intelligent, autonomous robots:the gap between the ability to reason about the world and the ability to sense the world and translate that sensory data into a symbolic model.
This ability is what enables living systems to look at the world and perceive the things in it. In addition, intelligent living systems can extrapolate from their mental models and predict the effects of their actions in the real world. The authors call this bi-directional mapping of sensor data to symbols and symbolic manipulation onto real world effects reification. After exploring the gulf between bottom-up and top-down approaches to autonomous robotics, the book develops the concepts of reification from biologically based premises"

It seems like the perfect book. Biology, computers, building things...

 

[lerchmo]

Reification is a concept that is explained from a neurobiological perspective in "On Intelligence" By Jeff Hawkins. His hypothesis is that the fundamental circuit of the brain is the a hierarchy of invariant representations of reality. So a house is composed of a number of invariant representations lower down in the hierarchy, Door, Window, Glass, Wood Line, Curve ect.

He calls his model the Memory Prediction framework and basically memories are formed as hierarchies and prediction flows down the hierarchy. From House down to wood ect.

I think the Blue Brain project is an interesting approach, I also think Jeff Hawkins approach with HTM / Numenta is interesting.

I think the problem being solved is fundamentally different in many of these situations.

For instance Blue Brain is trying to make a direct translation from a biological brain to software. And Numenta is trying to create an algorithm that works within the memory prediction framework.

While I am not an expert in low level computing and electrical engineering. I feel like mapping a brain to current computer technology is extraordinarily inefficient.

I am interested in projects that are trying to solve the problem at a physical level like: http://web.mit.edu/newsoffice/2009/ai-overview.html

Also I think many of these projects focus solely on the neo-cortex and don't regard the importance of imotion and the animal brain in directing and guiding learning and recall.