Artificial Intelligence: Thinking Big - KQED QUEST

Though computers have gotten faster, smaller and more versatile, it's still a big challenge to get them to demonstrate intelligent behaviors. Will machines like robots ever match -- or perhaps even exceed -- the capabilities of the human brain? QUEST meets a robot that in ten years time could take care of tasks around the house that most of us would rather not do.

Category:

Science & Technology

Tags:

• kqed
• quest
• pbs
• science
• robots
• AI
• artificial
• intelligence
• robot
• engineering

[grozmo1]

It's one thing to recognize a cup (various shapes/sizes) and another to 'understand' the purpose of a cup, determine other objects which could be used for the same purpose (eg A bowl is able to hold liquid) and even further be able to manipulate an unrelated object for the same purpose (eg Fold a leaf a certain way and it is able to hold liquid too).

Its more about purpose and intent than specific objects. A truly intelligent robot could see the world in this way - like Macguyver! haha

[MachineFight]

At 4:48 I thought it was going to use the PC. :D

[PoorMansMind]

5:17 that's exactly how i found this video. :D

[97448able]

I'm located in San Luis Obispo, California (moving soon). It could be age related (been there, done that); however, I have never been in a community with such a lack of Cultural Identity and originality. Other than some Robotics, there really is no need for spending: time, resources, and money on Artificial Intelligence...just come here.

Gurdjief's notion that the vast majority of humanity is nothing more than unconscious sleep walking "machines" is obvious here. The Bible etc, echo this notion.

[97448able]

@97448able While it is true that Sanskrit Grammar is highly analytical, precise, and scientific, The Sanskrit language itself is not, and is quite open to interpretation. Panini's Grammar the Ashtadhyayi is cited by some as a model for AI (Rick Briggs from NASA 1985!) the rules themselves, a complex metalanguage are great for analysis, but don't convey much in the way of meaning. These articles are great for learning about Sanskrit, but not AI. As you say: "AI lies in the hardware.

[97448able]

@chroniclerofthe70s Thank you food for thought. I thought the NSA had the largest, most, powerful supercomputer? oh well, I'm learning.

I'm a Student of Sanskrit and have read several of those articles pertaining to using Sanskrit as a model for AI and Computational Linguistics. True Panini's Grammar system might require a supercomputer for proper use...it is that monumental! However, for AI I'm quite skeptical would require quite a leap of faith...I will remain open to that.

[chroniclerofthe70s]

@TwistedLemniscate The emergence of intelligence of living systems is based on intrinsic combinatorial nature of intrinsic nature of interacting biomolecular processes with structure, densities, charge distributions, etc. In my opinion, AI lies in the hardware. A human trying to guess every possible scenario in code is not likely to achieve intelligence. Nature evolved intelligence with bio-organic molecules. Humans could also achieve this using design molecules.

[chroniclerofthe70s]

@97448able The K supercomputer has immense parellel processing power and yet it probably couldn't understand a human joke. The researcher is assumming the system is learning like a living system. The AI system might be learning patterns not recognized by the average human. In my opinion, a truly intelligent AI will be composed synthetic materials which mimic the biochemical and other physical processes of living systems.

[chroniclerofthe70s]

@benjamincrouzier I believe you are describing apparent intelligence not conscious awareness intelligence. Even if an AI system passed the Turing test, it wouldn't confirm if the AI system understood the communication exchange in the way a conscious human might. For instance, a human can distinguish a joke from a serious statement by it's context relative to common knowledge and experiences. An AI which could get the joke would be intelligent with respect to the Turing test.

[wiseyeffect]

They are not showing the true progress that has been made. The robotic they have now are far superior than Terminator but as stated from another user "this isn't AI". Look into "Watson" if your looking for AI.

[benjamincrouzier]

Artificial intelligence has nothing to do with robots. If you have a question-answering system that passes the Turing test (ie, make you believe that you are chatting with a human), I would be considered intelligent, yet it has no mechanical arms.

[97448able]

Absolutely fascinating, food for thought. Especially the ability to endow these machines with "Extra" senses humans don't have, see: 03:35 Laser Scanner etc.

Achieve a high enough "sampling rate" these machines will eventually mimic (ape) humans.

Than again Artificial Intelligence has already been perfected...just look around, you are surrounded by clones: "monkey see, monkey do, lookey me, lookey you". Is it live;or, is it memorex? Same old shit, different day.