Brain capacity limits exponential online data growth

Feb 01, 2012

Scientists have found that the capacity of the human brain to process and record information - and not economic constraints - may constitute the dominant limiting factor for the overall growth of globally stored information. These findings have just been published in an article in EPJ B by Claudius Gros and colleagues from the Institute for Theoretical Physics at Goethe University Frankfurt in Germany.

The authors first looked at the distribution of 633 public internet files by plotting the number of videos, audio and image files against the size of the files. They gathered files which were produced by humans or intended for human use with the spider file search engine Findfiles.net. They chose to focus on files which are hosted on domains pointing from the Wikipedia and the directory dmoz.

Assuming that for data production are proportional to the amount of data produced, these costs should be driving the generation of information exponentially. However, the authors found that, in fact, economic costs were not the limiting factors for data production. The absence of exponential tails for the graph representing the number of files indicates this conclusion.

They found that underlying neurophysiological processes influence the brain's ability to handle information. For example, when a person produces an image and attributes a subjective value to it, for example, a given resolution, he or she is influenced by his or her perception of the quality of that image. Their perception of the amount of information gained when increasing the resolution of a low-quality image is substantially higher than when increasing the resolution of a high-quality photo by the same degree. This relation is known as the Weber-Fechner law.

The authors observed that file-size distributions obey this Weber-Fechner law. This means that the total amount of information cannot grow faster than our ability to digest or handle it.

Explore further: Thermoelectric power plants could offer economically competitive renewable energy

More information: Gros C., Kaczor G., Marković D., (2012) Neuropsychological constraints to human data production on a global scale, European Physical Journal B (EPJ B) 85: 28, DOI 10.1140/epjb/e2011-20581-3

Related Stories

Program does impressive file size reductions

Sep 18, 2009

We intuitively understand the value of being able to make things smaller without sacrificing performance. The endeavor produces smaller speakers with bigger sound and a host of portable electronic devices such as digital ...

Multimedia search without detours

Apr 02, 2010

Finding a particular song or video is often no easy matter. Manually assigned metadata may be incorrect, and the unpacking of compressed data can slow up the search. DIVAS, a multimedia search engine, uses digital fingerprints ...

Hide files within files for better data security

May 09, 2011

Steganography is a form of security through obscurity in which information is hidden within an unusual medium. An artist might paint a coded message into a portrait, for instance, or an author embed words in the text. A traditional ...

Digital Foci Ships Image Moments 6 Digiframe

Oct 06, 2008

The Digital Foci IMT-063 Image Moments 6 digiframe features a 5.7 inch 640 x 480 display with LED backlighting, two tone chrome finish, 450Mb of internal storage, support for memory cards, and USB 2.0 connectivity. ...

Mozy Pro Offers Backup for Business

Apr 03, 2007

Berkeley Data Systems today announced the general availability of Mozy Pro, a business-oriented version of the company's popular consumer-oriented Mozy Remote Backup.

Recommended for you

User comments : 25

Adjust slider to filter visible comments by rank

Display comments: newest first

axemaster
4.5 / 5 (8) Feb 01, 2012
Uh, I don't follow the conclusion. What kind of data are they talking about?
Their perception of the amount of information gained when increasing the resolution of a low-quality image is substantially higher than when increasing the resolution of a high-quality photo by the same degree.

Yeah, there's a really good reason for that. Increasing the resolution of an image is only useful up to a certain point because any information gained beyond that point is USELESS and just increases the file size. You see, there's this thing called "statistics" and "error", and "optimization of error"... For example, in particle physics, you only need a certain number of detections to get good data - going any further is typically counterproductive because the file size goes up exponentially.

Please tell me they knew this... please...
sstritt
3.4 / 5 (5) Feb 01, 2012
I get it that this may place a limit on individual file size, but you can't conclude that it limits total global information storage!
Just_some_guy
Feb 01, 2012
This comment has been removed by a moderator.
gmurphy
5 / 5 (6) Feb 01, 2012
@sstritt, I agree. I think this study omits the fact that humans are adept at repositioning their perspective so that previously complex phenomenon are abstracted under a few general principles, freeing up cognitive resources for additional comprehension. In addition, the complexity of an environments need not be fully 'grasped' by entities operating within that environment, order for these entities to function effectively. How many people fully understand how a calculator really works?, I would image very few yet the inherent complexity of this device is not an obstacle for it's utilisation due to how it's presented to the user.
Eikka
4 / 5 (4) Feb 01, 2012
One of the reasons is, that the end user devices are limited.

The typical computer monitor has a resolution of roughly 100 DPI, and a "High Definition" monitor has roughly 2 megapixels to view images with. Viewing an image of a better resolution just looks exactly the same, or potentially worse because of interpolation and aliasing errors.

Likewise, people complain when others use the PNG file format that allows one to save pictures without losing any information. The problem is, that the files become 4-5 times larger while the difference is imperceptible and the penalty is having to experience the dialup-effect all over again where you're watching an image load up line after line over a slow link.
Jonathan_Robin
3.7 / 5 (3) Feb 01, 2012
The study's authors seem to have "found that underlying" ... funding ... "processes influence their brains' ability to handle information. For example, when some" ... grant ... "produces an image and attributes a subjective value to it," ... the authors pre-define... "for example, a given resolution"
tlootno
5 / 5 (1) Feb 02, 2012
Why do you waste my time with this completely absurd, incomplete and absurd article? Please, read the article before approving it...
antialias_physorg
4.2 / 5 (5) Feb 02, 2012
I think you're all not getting what the article is stating:

If economics were the driving factor in file generation (i.e. if cost per byte were the limiting factor) then we'd see a different distribution of filesizes (i.e. we'd see files growing larger in step with Moore's law- whichis in a LINEAR fashion). Such a growth would be due to the users' desire to have the maximal resolution affordable.

But this is not observed.

Instead there is a non-linear relationship between resolution (size) and usefulness. Going from 640x40 to 800x600 confers a larger gain (in terms of information gained by the user's brain) than going from 1280x1024 to 1600x1200.

People will create files that are useful - and if usefulness tops out (or goes asymptotic) towards higher resolution then it's not Moore's law that determines how much information is generated by people.
Noumenon
Feb 02, 2012
This comment has been removed by a moderator.
rawa1
1 / 5 (4) Feb 02, 2012
What a crappy article!, Why do you waste my time with this completely absurd, incomplete and absurd article?
What to waste time with such crappy arguments? If you have some relevant objection, just present it and avoid of dummy labelling.
This means that the total amount of information cannot grow faster than our ability to digest or handle it.
I'm facing it all the time during presentation of AWT. Many people here cannot handle these Victorian era models of physics, simply because they've no brain capacity for it. They don't know many aspects of solitons, waves, gradients etc. behaviour, so that my illustrations are as nontransparent for these people, as the high level math of mainstream theories. It seriously limits the speed in which the new ideas are accepted, because the people simply cannot imagine their deeper motivations. For example many people refuse cold fusion just because they don't know about most of experimental results connected with this phenomena.
Eoprime
5 / 5 (3) Feb 02, 2012
rawa calippo Zephir ... AWT ... cold fusion .. bla


If your are not able to back up your AWT nonsense with some credible source I will go along with reporting every comment of you where the words AWT, Cold fusion or waterripplecmbrnoisewhatever are contained.
---
Keep science: Include references to the published scientific literature to support your statements. Pseudoscience comments (including non-mainstream theories) will be deleted
antialias_physorg
1 / 5 (1) Feb 02, 2012
(i.e. we'd see files growing larger in step with Moore's law- whichis in a LINEAR fashion)

Brainfeeze...exponential, of course.
rawa1
1 / 5 (1) Feb 02, 2012
@Eoprime: Cold fusion is supported with NASA officially and here you can find 30.000 of articles about it, many of them are peer-reviewed. http://www.lenr-canr.org The people like you will ignore them all - so why are you asking some references, after then?
The people like you are just demonstrate the limits of brain, which I mentioned above. If they cannot comprehend something, they will simply censor it - this is exactly the mechanism, in which the cold fusion has been delayed for years. The fact some individual refuses some theory means anything in context of the fact, 60% of Americans are refusing an evolution. If 60% of people would refuse the aether model or cold fusion, is it evidence of being wrong, after then? And vice-versa: if 40% of people supports Big Bang theory, is it a proof, this theory is correct? Your stance is just poorly masked tendency to censor every idea, which goes over your ability to comprehend it.
Moebius
2.3 / 5 (3) Feb 02, 2012
The authors observed that file-size distributions obey this Weber-Fechner law. This means that the total amount of information cannot grow faster than our ability to digest or handle it.


Then maybe we need the equivalent of lawyers to mediate the information. There were 40,000 new laws passed in the US last year. Evidently very few were removed. The tide of new laws doesn't seem to inhibit justice by information overload so it must be the lawyers superhuman abilities to sort through all those laws and deliver a just legal system as we all know it is.
antialias_physorg
2.3 / 5 (3) Feb 02, 2012
and deliver a just legal system

First semester law school (as reported to me by my brother who is now a lawyer): You get taught that 'legal' and 'just' are two different concepts.

The legal system is not concerned with what is just. Only with applying the law.

If you're asking for lawyers/judges to help justice triumph then you're barking up the wrong tree.
Tausch
2.5 / 5 (2) Feb 05, 2012
One third of a human life is sleep. We had the 'lights turned off and on' for us for very long time - from the time we came into existence. All life has this still.

Not all life sleeps. Or runs a risk of input overload.
Our vocabulary went beyond rocketry by adding new meaning to the word 'burn out'.

Nature provides infinite resolution. Even without sleep and enhanced senses, not all of Nature's details will be accessible to us.

The power of infinite resolution renders the word 'evolution' meaningless. We don't have the imagination to exist without the word 'evolution'. We do have the vocabulary for that eventuality though:
God.

And that word won't have the slightest taint of religion when our vocabulary abandons the use of the word 'evolution'.

As a student, law schools will always give you more than you can handle. Intentionally. You are forced to be selective.
bewertow
not rated yet Feb 05, 2012
great, more off topic crap from rawa/callipo.

Reported. I suggest anyone else who finds this sort of spam annoying also reports it so this guy can finally get banned.
that_guy
5 / 5 (2) Feb 07, 2012
I think you're all not getting what the article is stating:

If economics were the driving factor in file generation (i.e. if cost per byte were the limiting factor) then we'd see a different distribution of filesizes (i.e. we'd see files growing larger in step with Moore's law- whichis in a LINEAR fashion). Such a growth would be due to the users' desire to have the maximal resolution affordable.

But this is not observed.


I think you're right - But everyone else is arguing that it's a stupid argument to make that information will keep growing as a factor of economics. Even a cursory analysis would have shown that the study would have been much more useful if broadened.

Data mining is subject to diminishing returns (As data increases) to the point where even finding new ways to mine data requires too much creativity/human wages to justify the costs. We are arriving on that point.

Continued
that_guy
5 / 5 (2) Feb 07, 2012
There was a point when your 100MB drive held maybe a couple games and financial data, etc. Now your terabyte drive holds every picture you ever took at full resolution, numerous shows in HD, your entire music collection, games, pron, and whatever else. Making a bigger hard drive won't make you use it more.

There's a limit as an individual of useful data - even if you're a data collector like my GF.

If an entity has everything I've ever done stored, does having more storage make them use more storage? No.

Once the need for data storage is satisfied, the only thing that will substantially contribute to data growth is new applications which require higher data use. (HDTV or streaming movies, for example)

In fact, the brain's capacity has very little to do with data storage, as other commenters have pointed out.
Tausch
3 / 5 (2) Feb 07, 2012
Interesting take.
I don't even know what unit (of measure?)science uses to handle 'brain capacity'. Our 'storage' falls short of a technical description.
Humans aren't ready for total recall.
antialias_physorg
3 / 5 (2) Feb 08, 2012
Making a bigger hard drive won't make you use it more.

Agreed. Unless new modalities come along which require vastly more disc space (full holographic recordings, brain state abstracts, etc. ). So for the consumer market I also see a limit in the needed capacity (with streaming and cloud storage even less!).

In commercial enterprises the hunger for ever more data storage is likely to continue for some time.

As you note: data mining becomes costly/tricky, so I guess the trend will go towards relating data as you store it (which is done in a rudmentary fashion in relational databases already).

The relations could potentialy take up much more space than the data itself if the connections go beyond the immediate domain of the data in question.

There are orders of magnitude more connections in the brain than there are neurons. (Though this is an imperfect analogy. The data is stored also in the weight of the connections in the brain)
Eikka
not rated yet Feb 08, 2012
In fact, the brain's capacity has very little to do with data storage, as other commenters have pointed out.


One could argue for example, that digital cameras have stopped at 12 megapixels because the human brain doesn't need any more detail in the picture. But, for example, an 8x10 print of a photograph with a typical consumer grade printer is 7.2 Mpix even though the printer claims ridiculous print resolutions like 1200 DPI. (Where the trick is, that one "dot" is actually not a full pixel)

If you took an 8x10 print and made it so sharp that your eyes can't possibly recieve any more information from the picture at 1' distance, you'd need about 15 Megapixels, and the source image must be larger at ~25 Mpix to allow for cropping and noise reduction by resampling etc.

So, the consumer camera resolution cap isn't because we can't see any better, but because we can't print any better. The printers are not good enough, the paper is too coarse, and finally, nobody cares enough.
Eikka
3 / 5 (2) Feb 08, 2012
And also because, if you keep increasing the number of pixels in a camera sensor but don't increase the physical size of it, the sensitivity goes down and noise goes up, so having more megapixels in a camera leads to all sorts of practical problems like how to get enough light to the sensor.
kochevnik
1 / 5 (1) Feb 08, 2012
No proof at all that memory even exists in the brain. Of course neural nets are a necessary requisite for recollection. So are a thousand other things. But without nested electromagnetic vortex entrainment you are brain dead, even if your synapses haven't snapped.

People always compare what they don't know, like the brain, to whatever they do, like computers. The metaphor is fallacious.

Usually the people who talk about "brain capacity" are hinting at eugenics.
antialias_physorg
not rated yet Feb 08, 2012
So, the consumer camera resolution cap isn't because we can't see any better, but because we can't print any better.

However if people want to zoom in (or produce larger prints) then the 'need' for more megapixels is again there.
the only thing that is realy limited is how many bits you need to represent color per pixel. The threshold for color perception isn't dependent on resolution.
that_guy
not rated yet Feb 08, 2012
@Eikka - I think you're off on some of the megapixel/camera issues.

The factors that go into that are independent of this discussion.

1. (As you stated) At standard sensor size, you gain 0 increase in resolution by going over 12 megapixels. This is due to physics of light wavelength. There are bigger sensors on some DSLRs and scientific cameras (And higher resolutions) but the increase in cost is substantial.

2. Typical commercial printing is 300 DPI, not 1200 dpi. 1200 dpi is generally only for photos that are meant to be viewed close up. Your printer at home may claim 1200 dpi, but it's effective resolution when you take accuracy into account is closer to 600.

The iPhone retina display is 326dpi.

3. The human field of focus is only about 15 Megapixels. so for practical display purposes, going higher provides much lower returns. (peripheral vision is much larger, but very not useful regarding static images.)
that_guy
not rated yet Feb 08, 2012
But in the end, these are mostly technical limitations that will almost definitely be overcome in the next 10 years. Once that happens, the rate of photographic data will only be a function of population and culture - At least for the developed world.

While photographers will prefer Raw images at higher resolutions, I think that will top out at 50-100 MP, and consumer cameras will likely never average over 20MP.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.