RSS 2.0 Feed

» Welcome Guest Log In :: Register

Pages: (622) < ... 601 602 603 604 605 [606] 607 608 609 610 611 ... >   
  Topic: A Separate Thread for Gary Gaulin, As big as the poop that does not look< Next Oldest | Next Newest >  
fnxtr



Posts: 3504
Joined: June 2006

(Permalink) Posted: Oct. 27 2017,17:24   

Same as it ever was.

--------------
"[A] book said there were 5 trillion witnesses. Who am I supposed to believe, 5 trillion witnesses or you? That shit's, like, ironclad. " -- stevestory

"Wow, you must be retarded. I said that CO2 does not trap heat. If it did then it would not cool down at night."  Joe G

  
Henry J



Posts: 5786
Joined: Mar. 2005

(Permalink) Posted: Oct. 27 2017,23:32   

Quote (N.Wells @ Oct. 27 2017,16:15)
Quote
OK, list the "predictions" your so-called "model" predicts, I bet you can't.


Legitimate predictions, logically entailed by the model, of course, not your usual "hypofeces".

Yeah, but that would require having a model that has some relevance to the claims.

  
coldfirephoenix



Posts: 62
Joined: Sep. 2017

(Permalink) Posted: Oct. 28 2017,08:53   

Quote (Henry J @ Oct. 27 2017,23:32)
Quote (N.Wells @ Oct. 27 2017,16:15)
Quote
OK, list the "predictions" your so-called "model" predicts, I bet you can't.


Legitimate predictions, logically entailed by the model, of course, not your usual "hypofeces".

Yeah, but that would require having a model that has some relevance to the claims.

Or having a model at all, not just a bunch of unhinged ravings.

  
Alan Fox



Posts: 1556
Joined: Aug. 2005

(Permalink) Posted: Oct. 28 2017,13:14   

[quote=GaryGaulin,Oct. 26 2017,19:24][/quote]
 
Quote
Wesley must be loving the help from all his friends who jump right in to help throw their malicious words around.
Bear in mind there may be others who hold back because ... well, in my case, because I don't see Wesley needs any help.
 
Quote
This is all very very creepy.
Certainly is!  
Quote
I'm expected to forever answer to semantic issues like what I consider to be "old junk" and what I think is "new junk" and whether "junk" you already played with and grew bored of is still "good junk" or just plain "bad junk". I have no interest in ponying up anything that only wastes my time answering to biased jackasses.
Then move on.
 
Quote
My job is to develop a MODEL that is still doing amazingly well making cognitive science related predictions that result in increasingly complex behavior in shock zone and other tests.

Does it pay well?  
Quote
I don't need anything more than that. At this stage I very much have the right to expect something better from critics, who as you can see are still very busy making excuses for not having anything at all.
I'm not sure why you <think you>* have that right. You certainly have the right to your own beliefs. But if you present ideas in a public forum, you should expect criticism. Don't like criticism? There's a simple solution. Move on.

I think this might be my second comment in this thread, Gary. I previously wished you happy birthday. I genuinely wish you good luck in the future... but I counsel you -- move on.

ETA*

  
GaryGaulin



Posts: 5385
Joined: Oct. 2012

(Permalink) Posted: Oct. 29 2017,03:18   

It is assumed that there is biological evidence galore for what is now academically popular, when in reality the evidence indicates that the currently accepted AI related methodologies are too loosely based on biology, very incomplete:

www.businessinsider.com/facebooks-ai-boss-in-terms-of-general-intelligence-were-not-even-close-to-a-rat-2017-10/?r=US&IR=T

The most modern evidence is now showing a process I have been experimenting with. I recently found this groundbreaking information having been recommended in a Reddit neuroscience forum:

Grid Cells and Place Cells: An Integrated View of their Navigational and Memory Function
www.ncbi.nlm.nih.gov/pmc/articles/PMC4679502/

The most modern evidence clearly enough shows that my critics are now the ones who are in serious trouble in regards to not having solid evidence to back up their claims. Even the AI community is aware that some of the now old-school stuff just isn't working as well as was hoped for.

Critics who do not even represent the experienced opinions from respected experts have no right to remain immune from meeting the same demands they make of others. The "delusional" are thus those who wrongly assumed that the evidence from other models is somehow on their side, when the opposite is now true.

What I'm up against is something that (for the sake of science) needs to be debunked. And I Won't Back Down.

Maybe I'm just a sucker for punishment. Perhaps for sake of my sanity I should take your advice. I now feel guilty for last night missing a WMAS Halloween Ball but with there being no money in this sort of science work we live without a car and other things that force us to live like hermits with the shame. I do though know what's right, and the for-real power of science that's on my side makes it easy enough to force the trash talkers to have to move on, instead. So I'll get back to work on the new ID Lab, which is turning out to be a RAM where all Data elements interact in a way that makes (what was there before in code using "Rnd") a best guess mechanism by more or less taking the randomness out of single point atmospheric/cosmic noise using multiple points to spatially map out the parts of the universe that generated the traveling waves. This leaves the "Confidence" requirement, which in this case has a 3.5/6 ratio indicative of network places most efficiently passing waves throughout a hexagonally connected geometry. Evidence of that existing in real brains has already been shown. For example:



At the moment the readings start off only a few percent too low, and go off a little more after getting shocks and hitting the wall enough times to become a little overwhhelmed but that is not all bad. Achieving perfection adds handshaking with two partnering places of six. In that case I'm using biological evidence to figure out behavioral details, which when accounted for eliminate a freezing up problem in the network that is caused by not having such a handshaking pattern. This can be seen in the video for 6.1 where the network places are noticeably still, yet do a great job keeping itself out of harm's way. What I had for signals helps show why it's a more than one way works sort of thing where achieving perfection will only result in a better guesser, not a whole new animal.

I'm tempted to again make each place in the network an autonomous bot that figures out the signal details. I earlier had some success, but needed to get where I'm now at with the code for that to become easier to include. In what I now have the two halves of the cycle already average out to an improved resolution that includes in between angles. Places just need a sense of when they are in perfect signal harmony with neighbors. I now think I know what the confidence level adjustment most simply needs for an OK signal.

The model I'm working on is still doing remarkably fine meeting the demands of modern neuroscience. The ones now in trouble lack brain wave action in a mapped out memory area and are missing an indicative signal pairing ratio somewhere in its numbers. It's the critics who have nothing.

There is nothing out of line in how I related the model to (whole cell behavior) cellular and (through replication of genetic memory) molecular level systems. What systematically applies to the top level may very well apply to levels from which it was emergent from. This makes it scientifically pointless for critics to take issue with a phrase like "molecular (level) intelligence" as though it's some kind of God in the gaps entity. I feel that it's best for others to understand why those who do so are scientifically annoying bung-holes, who are supposed to be the experts in testing for the same systematics being used at the biochemical level. Instead of my bowing down to common bully tactics from forum emperors with no clothes I would much rather thank you for your good advice that only figures I just didn't take by explaining all the above, then move on by in good conscious getting back to work on what is very clearly no delusion. It's then the best of both worlds where I can finally get all the uncalled for crap out of my mind and focus on the programming work. That's what the people who really matter are expecting from me. Moving on is not letting them down on account of what goes on in this thread. Your reply made it easy to do so, without my having to vanish from the thread in defeat. I just need to put the distractions completely out of my thoughts, look forward to better times ahead.

--------------
The theory of intelligent design holds that certain features of the universe and of living things are best explained by an intelligent cause, not an undirected process such as natural selection.

   
N.Wells



Posts: 1836
Joined: Oct. 2005

(Permalink) Posted: Oct. 29 2017,07:10   

If you think you have a working model of navigation via grid and place cells in a hippocampus, say that and build on it, but don't:
a) claim that you are modelling bugs/arthropods, which don't have hippocampi
b) claim that you are saying something about the origin of intelligence
c) claim that your work somehow supports trinities
d) claim that your work somehow shows conceptual shortcomings in natural selection
e) abuse standard definitions of "intelligence", and fail to provide logically valid redefinitions for "intelligence" that would allow you to use the word in non-standard ways
f) claim that emergence of intelligence somehow constitutes support for "intelligent design"
g) make unsupported claims about "molecular intelligence",
h) confuse your arguments with not only irrelevant but also incorrect assertions about crocodiles, salmon, and so forth and so on.

  
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Oct. 29 2017,15:39   

Animating research findings is not itself research.

--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
GaryGaulin



Posts: 5385
Joined: Oct. 2012

(Permalink) Posted: Oct. 29 2017,19:12   

Quote (Wesley R. Elsberry @ Oct. 29 2017,15:39)
Animating research findings is not itself research.

I'm modeling the neural spatial reasoning process neuroscientific research has been documenting. But I can understand why some need to make believe that's not true by calling it "Animating research" in order to suggest I'm making cartoons.

--------------
The theory of intelligent design holds that certain features of the universe and of living things are best explained by an intelligent cause, not an undirected process such as natural selection.

   
ChemiCat



Posts: 532
Joined: Nov. 2013

(Permalink) Posted: Oct. 30 2017,03:20   

Quote
I'm modeling the neural spatial reasoning process neuroscientific research has been documenting.


No you are not. You are, as Wesley says, animating a poor imitation of Pacman. There is no cognitive modelling involved especially as your insect has a hippocampus. This alone eliminates any scientific value of your "model".

Perhaps you would care to name any cognitive scientist who finds any use at all in your "research.

 
Quote
But I can understand why some need to make believe that's not true by calling it "Animating research" in order to suggest I'm making cartoons.


If you "understand why..." perhaps you can start answering the backlog of questions about your "research". Unless and until you answer your critics then your "model" is totally irrelevant to the science of cognition.

  
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Oct. 30 2017,07:34   

Quote (GaryGaulin @ Oct. 29 2017,19:12)
     
Quote (Wesley R. Elsberry @ Oct. 29 2017,15:39)
Animating research findings is not itself research.

I'm modeling the neural spatial reasoning process neuroscientific research has been documenting. But I can understand why some need to make believe that's not true by calling it "Animating research" in order to suggest I'm making cartoons.

Gary "Let's pretend!" Gaulin accuses someone else of make-believe. The projection is strong in this one.

Last time I looked, there was nothing remotely "neural" in Gary's code, and Gary specifically disavowed actually needing to do anything that would be an actual implementation of a neural network.

Gary is not, by his own admission, doing neural modeling in his PSC code. His statement of what he thinks he is doing is a perfect rationalization of an animator who does not think of himself as an animator.

Recapping "Let's pretend!"...

 
Quote

Gary does have something. It is his usual "Let's pretend!" routine, where anything he says is allowed to stand in for work he can't be bothered to accommodate. So integer values stand in for floating point, arrays stand in for neural associative memory, and nothing at all stands in for Heiserman's gamma level of robotic generalization. But in "Let's pretend!" land, this somehow is cutting-edge stuff.

Science doesn't value pretense, though, and keeps insisting on evidence via testing.


N-Bugs don't get there.

--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
GaryGaulin



Posts: 5385
Joined: Oct. 2012

(Permalink) Posted: Oct. 30 2017,17:06   

Quote (Wesley R. Elsberry @ Oct. 30 2017,07:34)
Last time I looked, there was nothing remotely "neural" in Gary's code, and Gary specifically disavowed actually needing to do anything that would be an actual implementation of a neural network.

Wesley has been complaining because I used the "Machine Intelligence" methodology I learned from David Heiserman, instead of the "Artificial Intelligence" neural network modeling technique that has since led others down a long road to the dead end I already knew about (from long ago having reached it that way too). This exactly:

www.businessinsider.com/facebooks-ai-boss-in-terms-of-general-intelligence-were-not-even-close-to-a-rat-2017-10/?r=US&IR=T

I'm none the less successfully modeling the behavior of populations of neurons, while my critics just throw defamatory insults.

--------------
The theory of intelligent design holds that certain features of the universe and of living things are best explained by an intelligent cause, not an undirected process such as natural selection.

   
ChemiCat



Posts: 532
Joined: Nov. 2013

(Permalink) Posted: Oct. 31 2017,03:50   

Quote
I'm none the less successfully modeling the behavior of populations of neurons, while my critics just throw defamatory insults.


HaHaHaHaHa---- Oh! Bloody Hell, you are serious---- HaHaHa.

It must be Manic Monday!

Bangles

  
Henry J



Posts: 5786
Joined: Mar. 2005

(Permalink) Posted: Oct. 31 2017,09:47   

Quote (ChemiCat @ Oct. 31 2017,02:50)
Quote
I'm none the less successfully modeling the behavior of populations of neurons, while my critics just throw defamatory insults.


HaHaHaHaHa---- Oh! Bloody Hell, you are serious---- HaHaHa.

It must be Manic Monday!

Bangles

Not here. Here it was 2:50 am Tuesday when you posted that. ;)

  
coldfirephoenix



Posts: 62
Joined: Sep. 2017

(Permalink) Posted: Oct. 31 2017,09:58   

Quote
I'm none the less successfully modeling the behavior of populations of neurons, while my critics just throw defamatory insults.



No and no. No, this is not what you are doing, stop playing make-believe in your delusional fantasyworld, and no, that is not what your "critics" are doing. Your "critics" are asking you incredibly basic questions and pointing out unbelievably easy-to-spot fatal flaws in your gibberish. You can't answer or address them.

  
ChemiCat



Posts: 532
Joined: Nov. 2013

(Permalink) Posted: Oct. 31 2017,10:42   

Quote
Not here. Here it was 2:50 am Tuesday when you posted that.


Gaulin's post was Monday.

My excuse is the time difference!

  
Henry J



Posts: 5786
Joined: Mar. 2005

(Permalink) Posted: Oct. 31 2017,22:22   

Oh, you were referring to his post time, not yours? In that case, ignore my comment.

  
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Nov. 01 2017,00:08   

Quote (GaryGaulin @ Oct. 30 2017,17:06)
 
Quote (Wesley R. Elsberry @ Oct. 30 2017,07:34)
Last time I looked, there was nothing remotely "neural" in Gary's code, and Gary specifically disavowed actually needing to do anything that would be an actual implementation of a neural network.

Wesley has been complaining because I used the "Machine Intelligence" methodology I learned from David Heiserman, instead of the "Artificial Intelligence" neural network modeling technique that has since led others down a long road to the dead end I already knew about (from long ago having reached it that way too). This exactly:

www.businessinsider.com/facebooks-ai-boss-in-terms-of-general-intelligence-were-not-even-close-to-a-rat-2017-10/?r=US&IR=T

I'm none the less successfully modeling the behavior of populations of neurons, while my critics just throw defamatory insults.

Again, animations are not themselves research. A system using operational processes different from the system of interest cannot be claimed to be anything more than an animation. And I'll note that Gary's summary of my complaints is false. Any review of the thread will show that I complained that Gary was using the reputation of Trehub to say that his PSC code was beyond criticism, yet Gary did not have any trace of Trehub neural models in his code.

I haven't said anything defamatory; I have made true observations and have plainly stated my opinions. Is Gary bugging Casey again?

--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
ChemiCat



Posts: 532
Joined: Nov. 2013

(Permalink) Posted: Nov. 01 2017,03:31   

Quote
Oh, you were referring to his post time, not yours? In that case, ignore my comment.


No problem. Gaulin's post was 17:06 Monday, with the time difference that is way past my bedtime! So I don't get to reply until the next morning. This happens all the time when I post on what is essentially an USA site.

  
GaryGaulin



Posts: 5385
Joined: Oct. 2012

(Permalink) Posted: Nov. 01 2017,06:24   

I did more searching for something new in Google Scholar, but still there is not enough solid information to make a neuron by neuron model of any part of the system. Even where it were possible the model would then only be representative of one type of animal and require a giant supercomputer to run the simulation in at least real time (no Turbo mode). It would be like writing a thousand lines of code to add 2 plus 2 instead of just simply writing "N=2+2", which is much clearer to the person trying to understand how it works. The answer should still be four, regardless of how much CPU time was wasted in the process of making such a simple computation.

With the function of even grid cells still unclear the model I now have has as much detail as is now possible. Going past that point is the wrong thing to do in a model intended to explain the fundumental basics of how something works. The critics are just delusional jackasses.

--------------
The theory of intelligent design holds that certain features of the universe and of living things are best explained by an intelligent cause, not an undirected process such as natural selection.

   
N.Wells



Posts: 1836
Joined: Oct. 2005

(Permalink) Posted: Nov. 01 2017,14:55   

Do
Snowflakes = Snowflakes + 1
Loop until Snowflakes >= Blizzard

Yay, I just modeled formation of blizzards.  Next up: watch me model hurricanes.

------

As Wesley observed, animation of something is not modelling it.

Computer graphics these days can produce exceptionally high quality animation of ripples of water, wind-blow water and grass, rippling fur, and the like, but these are not necessarily models of wind and water etc. in the sense of being scientific models of how they work so much as mathematical shortcuts toward producing realistic-looking effects.

As I've observed earlier, you need ground-truthing to make sure that what you are doing matches reality - from Wesley's comments, it seems that you are still a long way from doing that.

  
GaryGaulin



Posts: 5385
Joined: Oct. 2012

(Permalink) Posted: Nov. 01 2017,16:30   

It's payback time:

 
Quote
All of this behaviour-based robotics stuff was considered a bit kooky by mainstream researchers in the 1970s, who favored top-down strong AI. Why bother building little insect-level robots that puttered around on the floor? Machines needed to understand deep philosophical questions first. They needed to represent the entire world symbolically and reason about it like human brains. Only then would we be ready to put them on wheels or legs. So even though hobbyists almost immediately set to work building Buster clones, Heiserman was largely ignored elsewhere. But mainstream AI was already running into dead ends, entering what's now known as the AI Winter. And those Buster-building hobbyists were entering Universities and beginning to set the stage for a change in the direction of AI research. Before long, Rodney Brooks arrived on scene and coined the name 'subsumption architecture' to describe his own bottom-up, behaviour-based robots. Robotics and AI research were revitalized.

robots.net/article/3428.html

--------------
The theory of intelligent design holds that certain features of the universe and of living things are best explained by an intelligent cause, not an undirected process such as natural selection.

   
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Nov. 01 2017,18:14   

Quote (GaryGaulin @ Nov. 01 2017,16:30)
It's payback time:

     
Quote
All of this behaviour-based robotics stuff was considered a bit kooky by mainstream researchers in the 1970s, who favored top-down strong AI. Why bother building little insect-level robots that puttered around on the floor? Machines needed to understand deep philosophical questions first. They needed to represent the entire world symbolically and reason about it like human brains. Only then would we be ready to put them on wheels or legs. So even though hobbyists almost immediately set to work building Buster clones, Heiserman was largely ignored elsewhere. But mainstream AI was already running into dead ends, entering what's now known as the AI Winter. And those Buster-building hobbyists were entering Universities and beginning to set the stage for a change in the direction of AI research. Before long, Rodney Brooks arrived on scene and coined the name 'subsumption architecture' to describe his own bottom-up, behaviour-based robots. Robotics and AI research were revitalized.

robots.net/article/3428.html


Gary periodically links that article. The most recent instance where I responded to it was back in September, noting:

Quote

So, one person finds it "hard to imagine" that Brooks and others were uninfluenced, but doesn't have any current citations to show that they were. Gotcha.


That's not payback Gary is getting, it's memory loss.

--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Nov. 01 2017,23:18   

Quote (GaryGaulin @ Nov. 01 2017,06:24)
I did more searching for something new in Google Scholar, but still there is not enough solid information to make a neuron by neuron model of any part of the system. Even where it were possible the model would then only be representative of one type of animal and require a giant supercomputer to run the simulation in at least real time (no Turbo mode). It would be like writing a thousand lines of code to add 2 plus 2 instead of just simply writing "N=2+2", which is much clearer to the person trying to understand how it works. The answer should still be four, regardless of how much CPU time was wasted in the process of making such a simple computation.

With the function of even grid cells still unclear the model I now have has as much detail as is now possible. Going past that point is the wrong thing to do in a model intended to explain the fundumental basics of how something works. The critics are just delusional jackasses.

The world's most experienced projector speaks.

--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Nov. 01 2017,23:47   

The OpenWorm project took on a particular movement as their first behavior.

OpenWorm does suffer from a reluctance to animate things and declare "Let's pretend!" that different things are equivalent.


Quote

Why C. elegans?

In the field of neuroscience, one of the simplest organisms that are studied is Caenorhabditis elegans, or C. elegans for short. It only has 302 neurons, has a very consistent lifecycle, and is well studied. Its whole body has only 1000 cells total. With those 1000 cells it solves basic problems of feeding, mate-finding, predator and toxin avoidance using a nervous system driving muscles on a body in a complex world.

The cells in its body work together to produce its behavior. Instead of starting with the behavior and building a simple system to capture it, we are starting with making models of the individual cells and their interactions. If we do this correctly so that the cells act on each other as they do in the real organism, we will have a much more realistic model than we would get trying to go straight to the behavior.

This seems to us the only sensible starting point to creating a true biological simulation that captures enough details and has enough constraints to approximate real biology. Simulating a single cell that doesn't move (like a yeast cell) isn't going to provide us enough of a foundation to build up to more complex organisms by itself. If we can't accomplish a simulation at this humble scale, we'll never be able to do it at the massive scale of the human brain. The technology that would come out of this endeavor would be applicable to much more complex organisms down the road.


 
Quote

Validation

In order to know that we are making meaningful scientific progress, we need to validate the model using information from real worms. The movement validation project is working with an existing database of worm movement to make the critical comparisons.

The main goal of the Movement Validation team is to finish a test pipeline so the OpenWorm project can run a behavioural phenotyping of its virtual worm, using the same statistical tests the Schafer lab used on their real worm data.

More detailed information is available on the movement validation project page.


--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
GaryGaulin



Posts: 5385
Joined: Oct. 2012

(Permalink) Posted: Nov. 02 2017,06:18   

Here is how well the supposedly "ground truthed" AI neural networks that Wesley and others have been looking for and demanding from me are doing right now:

www.wired.com/story/googles-ai-wizard-unveils-a-new-twist-on-neural-networks/
Quote
GOOGLE’S AI WIZARD UNVEILS A NEW TWIST ON NEURAL NETWORKS

IF YOU WANT to blame someone for the hoopla around artificial intelligence, 69-year-old Google researcher Geoff Hinton is a good candidate.

The droll University of Toronto professor jolted the field onto a new trajectory in October 2012. With two grad students, Hinton showed that an unfashionable technology he’d championed for decades called artificial neural networks permitted a huge leap in machines’ ability to understand images. Within six months, all three researchers were on Google’s payroll. Today neural networks transcribe our speech, recognize our pets, and fight our trolls.
But Hinton now belittles the technology he helped bring to the world. “I think the way we’re doing computer vision is just wrong,” he says. “It works better than anything else at present but that doesn’t mean it’s right.”

In its place, Hinton has unveiled another “old” idea that might transform how computers see—and reshape AI. That’s important because computer vision is crucial to ideas such as self-driving cars, and having software that plays doctor.

Late last week, Hinton released two research papers that he says prove out an idea he’s been mulling for almost 40 years. “It’s made a lot of intuitive sense to me for a very long time, it just hasn’t worked well,” Hinton says. “We’ve finally got something that works well.”

Hinton’s new approach, known as capsule networks, is a twist on neural networks intended to make machines better able to understand the world through images or video. In one of the papers posted last week, Hinton’s capsule networks matched the accuracy of the best previous techniques on a standard test of how well software can learn to recognize handwritten digits.

In the second, capsule networks almost halved the best previous error rate on a test that challenges software to recognize toys such as trucks and cars from different angles. Hinton has been working on his new technique with two colleagues at Google’s Toronto office.

Capsule networks aim to remedy a weakness of today’s machine-learning systems that limits their effectiveness. Image-recognition software in use today by Google and others needs a large number of example photos to learn to reliably recognize objects in all kinds of situations. That’s because the software isn’t very good at generalizing what it learns to new scenarios, for example understanding that an object is the same when seen from a new viewpoint.

To teach a computer to recognize a cat from many angles, for example, could require thousands of photos covering a variety of perspectives. Human children don’t need such explicit and extensive training to learn to recognize a household pet.

Hinton’s idea for narrowing the gulf between the best AI systems and ordinary toddlers is to build a little more knowledge of the world into computer-vision software. Capsules—small groups of crude virtual neurons—are designed to track different parts of an object, such as a cat’s nose and ears, and their relative positions in space. A network of many capsules can use that awareness to understand when a new scene is in fact a different view of something it has seen before.

Hinton formed his intuition that vision systems need such an inbuilt sense of geometry in 1979, when he was trying to figure out how humans use mental imagery. He first laid out a preliminary design for capsule networks in 2011. The fuller picture released last week was long anticipated by researchers in the field. “Everyone has been waiting for it and looking for the next great leap from Geoff,” says Kyunghyun Cho, a professor at NYU who works on image recognition.


My work has likewise been towards modeling "small groups of crude virtual neurons". I focused on the motor navigation side of the system while Geoff Hinton was focusing on the sensory input side.

For my critics this is the worst-case scenario imaginable. I better prepare for an even more damaging bashing from them.

--------------
The theory of intelligent design holds that certain features of the universe and of living things are best explained by an intelligent cause, not an undirected process such as natural selection.

   
Henry J



Posts: 5786
Joined: Mar. 2005

(Permalink) Posted: Nov. 02 2017,09:15   

Quote (Wesley R. Elsberry @ Nov. 01 2017,22:18)
Quote (GaryGaulin @ Nov. 01 2017,06:24)
I did more searching for something new in Google Scholar, but still there is not enough solid information to make a neuron by neuron model of any part of the system. Even where it were possible the model would then only be representative of one type of animal and require a giant supercomputer to run the simulation in at least real time (no Turbo mode). It would be like writing a thousand lines of code to add 2 plus 2 instead of just simply writing "N=2+2", which is much clearer to the person trying to understand how it works. The answer should still be four, regardless of how much CPU time was wasted in the process of making such a simple computation.

With the function of even grid cells still unclear the model I now have has as much detail as is now possible. Going past that point is the wrong thing to do in a model intended to explain the fundumental basics of how something works. The critics are just delusional jackasses.

The world's most experienced projector speaks.

And yet he keeps practicing?

  
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Nov. 02 2017,10:34   

Gary's PSC code does not have neural anything, virtual or otherwise. Gary wants to play "Let's pretend!" that there is something with relevance to biological systems in his code, but he has been unable to show there is anything but a simulation of parts of a now antique robotic system going on there, written in a now legacy programming language. We've been through this before.

Geoff Hinton is still working in neural networks. Capsule networks are an instance of artificial neural network and aren't animations of final behaviors or "Let's pretend!" shortcuts. Gary's attempt to equate himself with Hinton is chuckle-worthy, for those who track things on the Baez crackpot index.

As to how well artificial neural systems are doing, even the Wired reporter said:

       
Quote

Today neural networks transcribe our speech, recognize our pets, and fight our trolls.


As Bernard Widrow noted in his 1987 plenary speech at the IEEE First International Conference on Neural Networks, his Adaline system, as represented in the Least Mean Square algorithm, was already deployed for adaptive noise cancellation and channel equalization tasks in *every modem in the world operating at 2400 baud or higher*.

Pretty good for stuff Gary wants to sneer at.

But Gary's sneering at artificial neural systems is an affectation he picked up only after I held his feet to the fire up-thread over this claim that he made:

       
Quote

That is part of the COGNITIVE MODEL that has NEURONS that they can make virtual ROBOTS with, like I do.


Once it was clear that Gary hadn't bothered to actually put anything like "NEURONS" in his PSC code, he suddenly went on this kick about how artificial neural systems (ANS) must be bad and inefficient compared to, as it turns out, just animating the heck out of things. Unless, of course, he thinks he can hitch a ride on the reputation of an ANS researcher like Hinton. Gary likely doesn't even recognize the essential contradictions he is making in how he approaches ANS research, with fawning and "this is my best bud now" sorts of talk for things he doesn't yet understand and dripping disdain for anything he thinks I like.

On talk.origins, someone who had made a complete hash of things complained that a critic was engaging in "character assassination". The response was that his character had committed suicide, and the critic was simply holding the autopsy. Similarly, Gary's perception of "damage" is all due to self-inflicted wounds. The faults are part and parcel of what Gary has eagerly encouraged others to look at. To me, Gary is kinda like those poor unfortunates who have no sensation of pain and are prone to harm themselves and have to rely on others to notice that they've set their hand on fire. Gary takes it a step further, with the equivalent of insulting anyone who tells him, "Your hand is on fire."

--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
ChemiCat



Posts: 532
Joined: Nov. 2013

(Permalink) Posted: Nov. 02 2017,11:59   

Quote
My work has likewise been towards modeling "small groups of crude virtual neurons". I focused on the motor navigation side of the system while Geoff Hinton was focusing on the sensory input side.


And what reaction has Professor Hinton given to your "model"? Has he praised it as a major breakthrough in cognitive science and robotics?

What! You haven't sent it to him? I wonder why not.

  
Jim_Wynne



Posts: 1208
Joined: June 2006

(Permalink) Posted: Nov. 03 2017,09:49   

Quote (ChemiCat @ Nov. 02 2017,11:59)
Quote
My work has likewise been towards modeling "small groups of crude virtual neurons". I focused on the motor navigation side of the system while Geoff Hinton was focusing on the sensory input side.


And what reaction has Professor Hinton given to your "model"? Has he praised it as a major breakthrough in cognitive science and robotics?

What! You haven't sent it to him? I wonder why not.

Several times in the past Gary has told of emailing someone far above himself in the cognition spectrum and having either received no response or a polite "That's nice" he reported back here that his theory was doing fine and that the person he emailed "had no problem with it."

--------------
Evolution is not about laws but about randomness on happanchance.--Robert Byers, at PT

  
k.e..



Posts: 5432
Joined: May 2007

(Permalink) Posted: Nov. 03 2017,10:00   

Quote (Jim_Wynne @ Nov. 03 2017,17:49)
Quote (ChemiCat @ Nov. 02 2017,11:59)
Quote
My work has likewise been towards modeling "small groups of crude virtual neurons". I focused on the motor navigation side of the system while Geoff Hinton was focusing on the sensory input side.


And what reaction has Professor Hinton given to your "model"? Has he praised it as a major breakthrough in cognitive science and robotics?

What! You haven't sent it to him? I wonder why not.

Several times in the past Gary has told of emailing someone far above himself in the cognition spectrum and having either received no response or a polite "That's nice" he reported back here that his theory was doing fine and that the person he emailed "had no problem with it."

Those people were probably worried Gary would show up on their doorstep with a fatwa.

--------------
"I get a strong breeze from my monitor every time k.e. puts on his clown DaveTard suit" dogdidit
"ID is deader than Lenny Flanks granmaws dildo batteries" Erasmus
"I'm busy studying scientist level science papers" Galloping Gary Gaulin

  
  18634 replies since Oct. 31 2012,02:32 < Next Oldest | Next Newest >  

Pages: (622) < ... 601 602 603 604 605 [606] 607 608 609 610 611 ... >   


Track this topic Email this topic Print this topic

[ Read the Board Rules ] | [Useful Links] | [Evolving Designs]