multi-dimensional physics thread (for Gisteron) ;-)

So what you are saying is that even the computer is aware of the image its supposed to replicate before it is even put into the brain? LOL how do you know the computer is not just displaying to the best of its ability what you asked it do do with the picutre. make a new picture that X percentage looks like the original out of a brain chemical.

What is that percentage anyway? Coincidence is 50 percent. So what is the percentage that is needed above 50 percent that validates the blury immage as the same enough to the original that it can be concluded to be the same image?

Fyxe wrote: So what you are saying is that even the computer is aware of the image its supposed to replicate before it is even put into the brain? LOL how do you know the computer is not just displaying to the best of its ability what you asked it do do with the picutre. make a new picture that X percentage looks like the original out of a brain chemical.

No, can't recall saying anything like that. Only that once the images from the brain scans are generated it can go on and compare them to the ones that were shown to the subjects.

Instead of presuming that everybody in the planet is a complete idiot who doesn't know what they are doing, why don't you actually read the paper? The only thing you are laughing at here is your own admittedly very dumb ideas about how to fake results. You are not laughing at what the study actually reports, because clearly you haven't even read what it says. I'm not regurgitating it in its entirety here for you, because that's why I linked it for your reading. You asked for evidence, now absorb it or move on already.

What is that percentage anyway?

Depends on the metric.

Coincidence is 50 percent.

Citation needed.

So what is the percentage that is needed above 50 percent that validates the blury immage as the same enough to the original that it can be concluded to be the same image?

Actually, none. Even an exact pixel by pixel match can technically still be coincidental. The conclusion is not that the image is "the same". The conclusion is that, given the sample size, the likelihood that matches be this high (on average) being coincidental is below a certain threshold. Nobody is concluding same-ness. The only thing you are giving to me is a display of how you spent no time either reading the paper or studying any methods of statistical data analysis. It's not difficult. You just can't be bothered.

you would love me to move on wouldnt you. ignore the obvious flaws in your thinking. Well I dont trust your precious sceintists as much as you seem to want to. and besides that I did read the stuff and even more, see below.

I dont need a citation for random chance. ever do statistics? its common knowledge that random chance is 50% if you flip a coin a million times you will get an average that it comes up heads half the time. besides that what metrics are they even using to compare images? are they picking out specific data points on each picture to compare? how many of those data points are being used? picture recognition softare? not sure?

what I posted below here is another arrticle. Im not sure how you are seeing anything that looks like the picture LOL. the duck looks like a spider, the panther looks like a rhino, the owl looks like a dog, the post and the window look like one of those kaleidoscope thingys. Your computer cant tell the difference between a duck and a spider!! not very smart after all.

Using algorithms to decode mental images isn’t new. Since 2011, researchers have recreated movie clips, photos, and even dream imagery by matching brain activity to activity recorded earlier when viewing images. But these methods all have their limits: Some deal only with narrow domains like face shape, and others can’t build an image from scratch—instead, they must select from preprogrammed images or categories like “person” or “bird.” This new work can generate recognizable images on the fly and even reproduce shapes that are not seen, but imagined.

To figure out what a person is seeing, researchers turned to functional magnetic resonance imaging (fMRI), which measures blood flow to the brain as a proxy for neural activity. They mapped out visual processing areas to a resolution of 2 millimeters as three people looked at more than 1000 images several times each. The goal was to consider just the activity in response to an image—like a leopard—and eventually have a computer paint an image that would produce nearly the same activity.

https://www.sciencemag.org/news/2018/01/mind-reading-algorithm-can-decode-pictures-your-head

All they are really doing here is programming an organic computer, not pulling images or ideas never imagined before out of a brain and they are not even using chemicals. they are using blood flow!! not very spectacular to me. we have been programming computers for a long time. I dont see this brain interface programming as much different.

Fyxe wrote: you would love me to move on wouldnt you. ignore the obvious flaws in your thinking.

No, I'd just like you to get to the questions you keep looking right past despite returning to page 13 again and again.

Well I dont trust your precious sceintists as much as you seem to want to.

I don't care. I don't trust the "sceintists" either. The data speaks for itself, all they do is present it to us lay people in a way we can make sense of it.

and besides that I did read the stuff and even more, see below.

Then why did you say that it was a picture-in, picture-out process, when the paper clearly stated that it also examined readouts from subjects that did not get a picture input? Did you read it and then chose to ignore that part? What's the difference? Might as well not have read it, for all the attention you paid it.

I dont need a citation for random chance. ever do statistics?

Yes, I did, actually.

its common knowledge that random chance is 50% if you flip a coin a million times you will get an average that it comes up heads half the time.

For a fair coin, yes. And, as you say, on average. The chance of getting half a million heads and half a million tails is actually rather slim. You are far, far more likely to get a slightly uneven distribution, though on average you'll be about as far and as often above the half mark about as often as you'll be below it. But that's the fair coin. Nobody said that images were coin-like. If we were talking about a binary picture, then matching any one pixel by chance would be as likely as 50%, but individual pixels are not recognizable features. Such a binary random noise image would technically match about half the target's pixels, fair enough. But that's as trivial as it is meaningless. If you know half the pixels were matched you can say nothing of what the image looks like. If you say half the features were matched, now you have a result worth talking about. Let's do a quick back of the envelope calculation for the likelihood of matching, say, 1% of features, shall we. I feel like doing some math because that's how I like to react to being talked down to by people who don't know what they are talking about. So here goes:

Let's say we have an image of one megapixel size. 1000px in width, 1000px in height.
Let's be generous and say that a "feature" is a region with a characteristic shape something between 10 and 30 pixels in size. For simplicity I'll use a benchmark size of 25px. We'll leave room for deviations later, don't worry.
Furthermore, let's say that half the pixels do not belong to or cannot be assigned to a feature. This, to be fair, is for fairness towards both sides, rather than generosity for one. Matches of featureless areas will not count either towards the total match, nor will mismatches count against it.
That leaves us with 500kpx of feature space, so something around 20 000 features, give or take a few hundred, maybe a thousand-ish.
So, what is the probability, by chance alone, to match 1% of those features? That's two hundred features, each comprising some 25 pixels, and matching each is a 50% chance...
So the probability to match one feature exactly is 2²⁵ to one against. That's 2.98x10^-6%, so 0.000003%. So far so good. Now let's consider how likely it is to match no less than 200 out of 20k.
Well, since the order of matching doesn't matter, it's the same as if we only considered one indexed order and the twenty first of that for the match. Alternative indexing schemes contribute to both the numerator and the denominator equally, thus they cancel out anyway. So therefore, the likelihood to match twenty features exactly is turns out to be 7.08x10^-1504%. That's 0.00...007% with some 1503 zeros between the decimal point and the first non-zero digit.
But that's not fair. That's an exact match. Okay then. How about we allow for... an entire fifth of any feature to be mismatched. Essentially, this is just means that for every feature, rather than having 25 stacked 50% chances, we only say that 20 have to be, and leave the other five to contribute as 100%, no matter if they match or not. Now the likelihood turns out to be 7.57x10^-1203% to get a 1% feature match with allowances for errors, for a binary image of only one megapixel in size, by chance alone. Of course I am overlooking things, I'm assuming that only one specific match counts and others do not. I'm assuming that pixels to match are selected in advance, which is only partly fair in that features comprise of neighboring pixels by definition. But we are talking about a single megapixel here. If we add all shifting, rotating, and flipping as options, we only improve the chances by something like six or seven orders of magnitude. I'll be generous and say that accounting for all such operations gives you back as many as twenty orders of magnitude. Heck, I'll even give you a binomial coefficient to throw in there, just in case stochastics was so long ago for me and I'm so tired to be overlooking something very crucial. That'll give you back just shy of 500 orders of magnitude. You'll still be left with half a grand of them short of the "common knowledge" that dictates the probability ought be 50%.
Now imagine what were to happen if instead of two colours there were 512 as in the case with RGB with only 3bits per channel. That's not even half the depth of what even the cheapest of our computer screens can display.

Of course this is assuming a by-pixel match, where features are defined as an arrangement of pixels in specific places. But people's vision is not binary. So what constitutes a feature would be defined differently, as would what constitutes a match. My point is, your "chance would be 50%" is far, far from obvious. Even a six sided dice already fails your intuition, as to throw a four, say, does not have a 50% chance, eventhough it is just about as random as the flip of a coin.
Part of the reason you and I both trust scientists is precisely because throwing their hands up in the air and condescendingly snark along the lines of "that's common knowledge^TM, I ain't gotta explain nuthin" is not how any scientific argument goes. Barely anything goes without saying, nothing is obvious, nothing "just makes sense" to be so. No matter how intuitive you find a number you suggest, if you want your work to be taken seriously, you don't get to just throw it out there and then neither do anything to justify it nor admit that it is but a working assumption.

besides that what metrics are they even using to compare images? are they picking out specific data points on each picture to compare? how many of those data points are being used? picture recognition softare? not sure?

Pretty sure they trained dedicated image recognition AI that learned from its own experience. As for how they evaluated the quality of the reconstructed image from the brain scans, feel free to consult the "Evaluation of Reconstruction Quality" subsection of the "Materials and Methods" section of the paper you assured me you read. It must be the one part of the paper that you overlooked when you read it. Or maybe you didn't read it. Judging by your output, I can't for the life of me tell the difference, and I barely skimmed over the paper myself, having no formal training in the subject to judge it with any fairness.

what I posted below here is another arrticle.

Yes, well, I linked the papers it was talking about, so nothing new to be gained there.

Im not sure how you are seeing anything that looks like the picture LOL. the duck looks like a spider, the panther looks like a rhino, the owl looks like a dog, the post and the window look like one of those kaleidoscope thingys. Your computer cant tell the difference between a duck and a spider!! not very smart after all.

The computer is not supposed to tell that difference. It doesn't even paint the images so much as reconstructing them from the signals it received from the brain. It then goes on to tell how well the images match, just so the researchers can put a number to it beyond just "well, it maybe kinda looks similar-ish or not" the way I imagine you would with your common knowledge. The computer doesn't care what any of it is "supposed to look like".
Also, I assume by "panther" you mean the leopard...

All they are really doing here is programming an organic computer, not pulling images or ideas never imagined before out of a brain and they are not even using chemicals. they are using blood flow!! not very spectacular to me. we have been programming computers for a long time. I dont see this brain interface programming as much different.

Irrelevant. I was only citing evidence for my assertion that yes, indeed, we can read (some) thoughts from physical signals. The "never imagined before" part only became an additional stiputlation of yours after I had met my burden of proof. It doesn't matter how impressed you are with the quality of that readout in the end. Nobody promised or owed you any impressions, only the facts. I'm not impressed with your reading skills either. Doesn't mean you do not in fact have any.

No, I'd just like you to get to the questions you keep looking right past

Exactly what questions were those again? I seemed to have missed them? can you point them out please?

Then why did you say that it was a picture-in, picture-out process

Because thats basically what it is. the person gets shown a picture of a duck and then thinks about it and then the computer generates a picture of a spider.


And thanks for all the maths stuff. I just skipped over that part so hope you enjoyed it. All it really boils down to is a smoke screen to cover the fact that you failed to prove we can read minds good enough to see ideas that have never been thought before. No 6 winged airplane will ever come out of the stuff they are doing because ideas dont really exist in this realm and cant be studied other than subjectively.

Fyxe wrote:

Then why did you say that it was a picture-in, picture-out process

Because thats basically what it is. the person gets shown a picture of a duck and then thinks about it and then the computer generates a picture of a spider.

What about the experiments where subjects were told to think of things and were not shown any pictures? You know, the other group they had for control. Are you ignoring it on purpose because you understand it's the opposite of what you claim the entire experiment was, or did you just "skip over that part" the same way you did "all the maths stuff" that showed how horribly wrong you were about "coincidence is 50 percent"?

And thanks for all the maths stuff.

You are most welcome. I hope you learned something reading through it.

I just skipped over that part so hope you enjoyed it.

Sure did. Sad that you ignored my detailed explanation as to how your "coincidence is 50 percent" is not obvious enough to be taken at face value. I bet you are going to say that I never showed you how you were incorrect, or that it was beside the point. I'll grant you the latter. You went back several times to page 13, quoted the message where I posed two questions for you, but are still continuing on this tangent even after promising you would address them.

All it really boils down to is a smoke screen to cover the fact that you failed to prove we can read minds good enough to see ideas that have never been thought before.

Never set out to prove anything of the sort. I only set out to provide evidence for the claim I made. And I did. And now you demand I provide proof of things I never claimed. I have no interest in meeting that demand.

No 6 winged airplane will ever come out of the stuff they are doing because ideas dont really exist in this realm and cant be studied other than subjectively.

I disagree. I'd wager that one day a six winged airplane will come out of the stuff they are doing. And you will either "skip over that part", or declare that it wasn't enough, because the six winged airplane you meant was obviously supposed to be solar powered and the one they read out of some subject's mind didn't even have adamantium armor. Because it's not about what was found, or what you said or what your interlocutor said. It's not about an honest exchange. It's about cheap victories for you.


It's this part that's most frustrating about you. Not that you happen to say nonsense. We all do that. But that you never consider that what you say is even up for discussion, let alone capable of being incorrect to any extent at all. You knowingly and purposefully ignore any and all challenge, and then pretend like none were raised to begin with, meanwhile never acknowledging any attempts that were made to meet your challenges, but instead shifting the goal post when ever everybody can see that the initial ones were met. Fyxe, you are the proverbial chicken the chess players were dumb enough to take on, the one that hops on the table, kicks over all the chess pieces, poops all over the board and then stands up proudly in the middle and declares victory.

It's not that you don't know what you are talking about... like ever. Most of us don't really know what we are talking about a lot of the time, too. I, for one, have no background in biology, let alone neuroscience. It took me all of five minutes to google the difference between touch perception and heat perception, but you knew better long before then and probably thought me an imbecile for even considering that you might be in error - as you were. The only thing I have on you is that I read and pay attention, and actually take the time to understand, before hopping on to compose a message after having forgetting who said what and lying as a result.

It's not that you don't know. I don't know. With you it's more than that. You won't know. Not for anything in the world. You are completely unwilling to learn, or to grow, or to compromise. You never engage in good faith, if we dare say that you engage at all. Almost the first thing you did upon coming to this forum is berate users for not being entirely onboard with your fantastical ideas. Aren't we calling ourselves Jedi, you asked. Isn't this a temple, a place of spiritual enlightenment, you wondered. Aren't we supposed to believe in magic powers, you implied. But we listened to you, we discussed with you, your philosophy and your science, because our Code tells us to expand our knowledge, and to judge reservedly, after hearing all positions in any case. Your code, if you have one, apparently does no such thing. Your code only insists that you have it all figured out, that you are above it all, and that whoever dares say something substantive toward you is beneath you, at least intellectually, possibly morally, too. Most certainly, so it must be telling you, they have nothing to say worth listening to.

Are we Jedi? Well, that is not a judgement for me to make. I have been thought of as one, I have been thought of as an adversary, too. In the end, for better or for worse, I am running out of chances to be handing out to you, to show to all here that you have something to contribute from whence all of us could learn. It appears, however, that my efforts are in vain. It appears, alas, that you are, properly, thoroughly, incorrigible. A lost cause.

I wish you well, Fyxe, where ever "the Force" takes you. Perhaps one day it may blast that solid gold helmet off your head so your sight extend beyond your nose's length. Let's hope it won't be an all too painful blast, should it ever come.

What about the experiments where subjects were told to think of things and were not shown any pictures?

The pictures were still shown to somebody sometime somewhere so they could be mapped by the computer. What it boils down to is that the computer cant pick anything out of a brain it does not previously know about. If the computer has mapped what it looks like for a human brain to think of a duck the computer spits out a spider, no matter who it is because it knows what a duck/spider looks like. but for example if someone thinks of a 6 winged plane the computer will never be able to spit out a sled or a tractor or whaterever else the computer things is a 6 winged plane ever because it has never mapped it and doesnt know about it. Its not reading minds, its simply regenerating previous informaiton its was programed to see.

where I posed two questions for you

Oh thats right I forgot you wanted me to answer some question. Hey can you sent me a link to them or something. I cant seem to find them.




So by your last comments are you saying you are abandoning our conversation? well thats fine with me but let me ask you one question. How does it feel when you are given what you dish out here typically? you are the most infuriating and condescending entity I have ever encountered here. Maybe we are in good company and you have finally just met your match in the negative traits you have assigned to me. Calling people liars and weasels is NOT the way to go about having a conversation. So I encourage you, if you are done, to at least take a bit of time and look inside yourself as well and judge that in the same way you judge me. And in that be brutally honest and stop lying to yourself that you are really just here in this thread trying to prove to the world that I am worthy of listening to.