And that is going to be a really big problem if Human level AI is based upon such networks.
And that is going to be a really big problem if Human level AI is based upon such networks.
-- Dirk http://www.neopax.com/technomage/ - My new book - Magick and Technology
el.
fle
,ce
dry
Better: swing a baseball bat at it. Then try the same with a fly, and see who wins.
My money's on the fly.
-- Cheers, James Arthur
Sounds hokey to me. Even if it works, they don't understand how it works.
"they may someday be used to fly small unmanned airplanes" which, I guess, is 10 years away.
How and where does our brain store images and words? How is vision encoded in the optic nerve? You won't answer questions like this.
John
Your ignorance is astounding. I'm no expert in neural research but even I know of things like this:
"ScienceDaily (Nov. 16, 2010) ? Researchers have developed an artificial retina that has the capacity to reproduce normal vision in mice. While other prosthetic strategies mainly increase the number of electrodes in an eye to capture more information, this study concentrated on incorporating the eye's neural "code" that converts pictures into signals the brain can understand. ... Using mice as subjects, the authors built two prosthetic systems: one with the code, one without. The researchers found the device with the code reconstructed more details. "Incorporating the code jumped the system's performance up to normal levels -- that is, there was enough information to reconstruct faces, newsprint, landscapes, essentially anything," Nirenberg said."
-- Dirk http://www.neopax.com/technomage/ - My new book - Magick and Technology
Then you two should be looking for dates on the 'short bus'. No surprise that you are both afraid of smart women.
-- You can't fix stupid. You can't even put a band-aid on it, because it's Teflon coated.
My kind of woman - Zoe Bell
-- Dirk http://www.neopax.com/technomage/ - My new book - Magick and Technology
The corollary of which is, no matter how great the hardware, the same crappy software will be run on it.
That should put it right after controlled nuclear fusion supplies all of our power.
Michael A. Terrell:
You should'nt. Your right hand is getting tired.
Oh, "artificial intelligence" isn't all that difficult - they've been playing chess for decades, and recently they've beenn touting some box that's going to compete on Jeopardy soon!
But they'll never achieve sentience until they can come up with a machine that can feel pain or joy.
But sentience is terribly misunderstood - an ameba is sentient - if you poke one, it will flinch! =:-O
Cheers! Rich
snipped-for-privacy@att.bizzzzzzzzzzzz:
Exactly. You can learn a lot about humans and neural networks in general (and, for that matter, any sort of biological or manufactured system) from their failure modes.
When a software made with an imperative language fails, it fails abruptly, inexplicably, "disgracefully". When a neural network fails, it may do so in a ridiculous way, but still sort of "familiar".
There are about 1.2 million neurons in each human optic nerve, 75K in a mouse. I wonder how many they connected to to get "normal" vision.
Which newspapers do mice prefer?
John
Not entirely true. At present Blue Brain is simulating one neocortical column. Those still have to be tied together in their thousands (millions?), so the s/w will not be identical but introduce higher level features impossible to simulate at present.
-- Dirk http://www.neopax.com/technomage/ - My new book - Magick and Technology
es
eare
neural
about
of
The
he
efar
an
et
.the
Not a catagory that John Larkin could identify reliably. He seems to get his information on anthropogenic global warming from denialist web- sites - mostly likely relayed through the right-wing media that he trusts - which tells us how reliable he is in picking real science from tendentious propaganda.
Perhaps not, but we do know that they are stored. For this kind of modelling, the actual mechanism isn't important.
The relevant academics do tend to talk about rapid-eye-movement (REM) sleep, and this does seem to be confined to mammals. At one stage monotreme mammals - the echidna and the platypus - were thought not to exhibit REM-sleep but it seems that this wasn't quite right.
-- Bill Sloman, Nijmegen
Given the redundancy involved in biological systems, maybe not a lot.
-- Dirk http://www.neopax.com/technomage/ - My new book - Magick and Technology
Actually, we do have a fair idea of how the information stored in the DNA gets transferred to messenger RNA. which in turn controls the synthesis of proteins. Eukaryotic pre-mRNA can be edited after it has been created from the DNA, before it is used to control the synthesis of one or more different proteins (depending on the editing). We are also starting to get an idea of how the varu=3Dious cells in the body know which kind of cell they are, and thus the sorts and quatities fo proteins they have to synthesise.
We don't know enough about it yet to have magic bullets to stop viruses - like the human immunodeficiency virus (HIV) that causes AIDS - but we are making progress.
Functional MRI is starting to tells us what bits do what. Molecular biology is starting to tells how the cells do what ever it is they do, and how some fo their control loops work.
Perhaps.
Or maybe less. Someoen who doesn't even know about messenger RNA isn't going to be the first to know wneh we do make a break-through.
Perhaps.
Any more than your capacity to design a circuit depends on knowing the fine details of the conduction band structure of the silicon and the other semi-conductors in the parts you use.
-- Bill Sloman, Nijmegen
Rich Grise:
Don't forget Eliza. But they were evidently in bad faith, I can't believe they were that stupid. But, you know, ads are vital for commerce.
Pain, mainly. Joy is an optional.
Joy is an optional.
neural
about
You are totally obsessed with this climate thing. It apparently leaves no room in your neurons for useful stuff, like electronics. Stick to things that are useless and untestable; you seem to be best at that.
Brilliant!
How can you do meaningful modeling of a system that you don't understand at any level? Well, if you can get a climate model to ape superficial past climate behavior, I suppose a brain model is no more difficult.
John
So, if a small number of optic-nerve neurons are adequate for full vision, why would nature use so many? The information theory implications are interesting here: full-motion encoding of landscapes and newspapers with a small number of low-rate nerve pulsers.
And which newspapers do mice prefer to read? You never answer the really important questions.
John
John Larkin:
I do. And a lot of other people.
Maybe. But Aasimov was not that stupid.
If they do sleep, maybe tey do remember some of the weird comeouts of their sleep, the usual outcome of messing with the weight of the connections of a neural network. Not the reminiscence, the ridiculous outcomes.
Weights? How can they be interpreted? No way.
Definitely.
I assume you are kidding.
Good analogy. Even if you're not that good with analogies.
You definitely lose at least 1/10**14 of your knowledge.
I can't go in detail, as you will not be able to, as nobody will ever. Is it you the one that is'nt able to comprehend that you cannot go into quantum mechanics' details?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.