It is known that in computer science there is no absolute randomness, and all data is predictable. There is also a grain of randomness. What if the grain is protein? There is also a data visualization site that claims that it is impossible to predict the outcome of a random image.
http://www.random-art.org/online/
I am not a bio computer scientist, but I would be interested to know how the proteins differ in which beautiful images are drawn from ugly images
TNEWKHKLKDMDCPWAYCETGEWEHRHFFPWSVIREIRVRYNKSVCCEDYFIDPYARGDN
RVATVHIQSWWDADWQIHGM
I found a protein with 80 amino acid residues. Which draws such a picture. such images are only 5%. The remaining 95% is pixel chaos
For example, this amino acid sequence
FTPDEIEMKLRNCDSLIWVNKGGPNFLHYVAHQHCPAWQIDEYDSEEPQKAHMMNHVTQE
DVRCQWKQHKFANGCDVFDY
That looks amazing, Steven Pletsch and a few others also experimented with images.
its almost like its working out angles; in some of them; off the original background image and deleting parts as it sees the code in the letters or adding them as it translates or mistranslates before communicating it.
Do you know of the Markov Chain ?
What amazing me most, to try and make sense of these... it's fascinating.
nature processes things using a formula and a self sustaining(repairing) system. at what exact "point" does this formula and protein we design connect with all the variables and parts we dont know, not to forget entropy and all the others laws we already know. So when every state is known; that wisdom brings a "bible" of sorts that is based on the language of the universe. The book of life. just wow really. It seems different entities unaware of themselves are giving a platform (so to speak) and in case of a god particle determining a point in time(as we know it) where something starts from somewhere, building from a point in a place that's connected to, commands, allows, welds, sews atoms, molecules and in the right place and the right time to sustain "life".. building from components around it(or drawn in by it) but in amounts to grow and sustain itself from the most compact area available. then for that thing to carry on propagating... mutating till its to a point where nature destroys it, or its "lifetime" decays to a point its unsustainable; runs out of fuel or reused by another species... maybe the art is hiding something we couldn't see right in front of us.
Have a great day, thanks for sharing.
For example, the data of this protein form the radii
CKRFYDHTCPGPAVMPVLGRAMVYRKGYFETSFEASEQKCCGNDLKCHGGCWTLFHNGHATNNTFAAIFYPLKDVHAYYE
HTKIYLCASNASVFSFCMTQINNCFCDHPVVNHEYSYYTHCTIYA
structured pixel noise
This protein gives amazing patterns.
WNYNVTCGQPWDTSTQPMCAYTWTVDRVGMETDSQSRNHAWIHKH
I do not know why. But finding beautiful proteins is hard to find. They are few, in fact. And they seem to take their place in the random
ICVGTKFSLATTVMRFVLKCGKTRDQMQAHPSVPSYIHFYNSYFY
here square structures turn into a rhombus
TSQGKDVCVRCFYSYRYYWITWHKCPFKNQMRMQMTFDCPFICRH
here again a circular image, we can assume that this is an atom or an elementary particle in a section or part of something else
LVIKNRDGIHMMPEPGQLSRMHVGCVAWHFKVGHWAKGHDHDVDK
there are some kind of spectral waves
The person who transmitted this image to me claims to have deciphered all the mathematical laws and laws of the universe.
The essence of the problem. Your data has limited accuracy.
The difference between 0.0001 and 0.000000001 is very large.
Below will be a link to full img
These are fractal patterns from simple to the most complex periodic fractions.
The result of many years of study of infinite periodic fractions.
In a few years, official scientific work will be released.
link to full img
Based on your website of "random-art.org", I may have discovered a correspondence between the meaning of American English words and the formula that the website uses.
I must be very lucky for this to happen.
Out of all the possible combinations of all of the pictures that can be created, I have found sections of English text that outputs a picture that partly resembles what the English text describes. The implication is that this seemingly "random" action on that website can be put into our advantage after all.
The first discovery
In "make your own" tab, with this picture name (referred to as "Text 1"):
itan is probably partially differentiated into distinct layers with a 3,400-kilometer (2,100 mi) rocky center.[29] This rocky center is surrounded by several layers composed of different crystalline forms of ice.[30] Its interior may still be hot enough for a liquid layer consisting of a "magma" composed of water and ammonia between the ice Ih crust and deeper ice layers made of high-pressure forms of ice. The presence of ammonia allows water to remain liquid even at a temperature as low as 176 K (−97 °C) (for eutectic mixture with water).[31] The Cassini probe discovered the evidence for the layered structure in the form of natural extremely-low-frequency radio waves in Titan's atmosphere. Titan's surface is thought to be a poor reflector of extremely-low-frequency radio waves, so they may instead be reflecting off the liquid–ice boundary of a subsurface ocean.[32] Surface features were observed by the Cassini spacecraft to systematically shift by up to 30 kilometers (19 mi) between October 2005 and May 2007, which suggests that the crust is decoupled from the interior, and provides additional evidence for an interior liquid layer.[33] Further supporting evidence for a liquid layer and ice shell decoupled from the solid core comes from the way the gravity field varies as Titan orbits Saturn.[34] Comparison of the gravity field with the RADAR-based topography observations[35] also suggests that the ice shell may be substantially rigid.[36][37]
( The source is a Wikipedia article on Titan(moon) at https://en.wikipedia.org/wiki/Titan_(moon) )
I just happened to choose the right paragraph and leave the "T" out . It is also very lucky that this paragraph was edited by a countless number of Wikipedia editors and with the right number of citations! Also the choice of vocabulary is very lucky too!
Key notes:
- Text 1 outputs 2 red concentric circles or different shades of red against a red background. This is similar to what the text is describing - the layers of the moon, Titan.
- Adding the "T" to the beginning of the text outputs a smooth color of green.
The second discovery
In "make your own" tab, with this picture name (referred to as "Text 2"):
Garden
The outputted picture resembles a garden. In the picture, there is a white diagonal streak that resembles a path. Also, there are gray circles that resemble moss-covered stones.
Finally, the rest of the picture is green, resembling grass.
Key notes:
- The text - "garden" yields a picture that resembles a bookshelf. The background is brown, so this could also resemble soil or perhaps a city of a sort.
All text has to be put in with exactly the same uppercase and lowercase letters.
I think that if we find a correlation between an amino acid sequence and a picture, then my hypothesis is that We may be able to find a correlation between an amino acid sequence and English text!
This would be a huge discovery with deciphering the amino acid sequence in Nature into English, of all languages!
I have been studying fractal theory for several years.
It is in this random art generation that the positive and negative colors are interesting. Further, you can already look at the curvature of the pattern. (It is also worthwhile to understand that all human conclusions are limited since a person has a limited brain)
For this project, there should be an AI that needs to understand correlation and dependence. between variables. And then perhaps we will understand in general all the principles of the universe, since the universe is simply data that can be visualized or predicted
Multiple names are likely to produce the same result.
Your “random-art” website may not be that random after all.
Within a week, 3 images were published that produced exactly the same painting. These painting’s names are D O R I M E, blue, Tam Test.
Also, painting bbbbbbb produces exactly the same painting as the other three referred to above, but was produced way earlier.
You may want to reconsider this website a little.
P.S. paintings bbbbbb and lue are not the same, so the “b” can not be subtracted from both sides of paintings “bbbbbbb” and “blue”, even though paintings “bbbbbbb” and “blue” are the same.
perhaps these symbols create the same grain and generation weight. There seems to be source codes for the project. If you know programming it is worth exploring.
The possibility that 3 symbols were published within a week that create the same grain and generation weight is highly unlikely according to the developer of the website.
In the about page of the website, the developer, Andrej Bauer, says: "It may happen that two different names determine the same picture, although this is rather unlikely due to the large number of all possible pictures."
The strange phenomenon seems to me as if there are more commonly-occurring pictures for the different pseudo-random seeds. This will be a problem for your experiment because multiple amino acid sequences may generate the same picture.
You solution with precise AI observation might solve this problem, as you have said that AI can detect things that humans can not, and I think that AI's observations might be a lot more sophisticated than ours.
Since this happened, we must consider the possibility of any future inaccuracies in the information we discover in the future.
I am not interested in how the website is programmed, but I do know how the website works according to the "about" page of the website, which is generating a pseudo-random seed out of the picture's name, and then generating an acyclic array from that seed.
The possibility of 3 pictures generating the same acyclic array is hardly fathomable to me, but this is something that I think we can put into our advantage. I think we can find how the website generates the pictures from observations of phenomena like this.
I have an idea on how to train an AI to detect differences between pictures better than us.
If the AI can detect the differences between all of these 3 pictures, then it shall be a viable AI because then our problem of pictures that look the same will be solved.
The three pictures are provided below. The names of the paintings are "Tam Test", "D O R I M E", and "blue" respectively. The source of these paintings are from random-art.org under the gallery tab.
Painting #1:
The Tam Test Painting
Painting #2:
The D O R I M E Painting
Painting #3:
The blue Painting
These are the same three paintings that were published within a week and all look the same, and started this problem.
I hope it helps!
The adding image icon does not work. The pop-up is nonfunctional and primitive. So I tried to use an HTML image tag, but that didn't work. Looks like Images can't be added to this post.
first one is "Tam Test' painting.
second is "D O R I M E" painting.
third is "blue" painting.
Sorry for the inconvenience.
In scenarios where Markov Chain modeling applies, sometimes Levy Flight simulation also lends itself to circular structuredness in protein patterns by teaching a neural network to simulate protein patterns. In some cases, fractal patterns will develop.
See what an interesting pattern!
Here is the protein sequence. And if we delete 1 character each and press enter, then always the colors or most of them will be solid colors.
QREVHIPYNVTDVKEMGGADNGFWCMYYPHGSQKQERDGELEGIPQVRPNEIIRSKPHLW
TELLTTDTPTYPGADVHFNQRIDCDKKPNWWAAEAIDLRRVNLLEREYQSYRYPFQQTEK
EGVKPYSCYVRWKDPYAYPYNEWM
That is, it is a protein of solid colors.
interesting discovery
I agree. Having the website painting generation program output solid colors, when 1 character is deleted each, sounds truly fascinating to me. This phenomenon can help us determine how the random art website works, and how it makes its paintings.
Also, I have a hypothesis for the second word in a picture's name.
If that word ends with a vowel, then the resulting painting is likely to be more tessellated and tapestry-like and having a repetitive design.
This is because in the "About" tab, the Demo Red and Demo Green do not have repetitive tessellated design because "Red" and "Green" end in consonants, not vowels. But since "Blue" ends with a vowel, the letter "e", the painting "Demo Blue" has a repetitive and tessellated design.
I wonder what do you think of my hypothesis? I think it is something good to be tested as a sort-of "sub-experiment".
Hi everyone! Due to continued spam on this post, we will be disabling comments for this thread. If you would like to continue this conversation, I would suggest creating a new thread. Thank you for your understanding.
For example, this protein draws a solid image.
HRYLPHCKMHGQTKIFWNWRCMYYFHMIGTAECNVFQCYFCVWCSSYWEHATSPKRLGQW VELCDREHWHTTLNWDTPGY
https://imgur.com/rcK5hSn
For example, this protein draws fractal patterns.
TWMMLDTCLQPVQHLGLPWALAHIPSYCTIRQIIKANCSYYFGCALPLVTQTPINMRMQS AIGADYSKNSTTVTHTIWEK
https://imgur.com/oj53Y10
Data from this protein shows circular structuredness. If you could find interesting structures by teaching a neural network. Or how to get knowledge of how it affects protein.
FRMMHDKCKYIFQHQSQWCPFEDNGSYWMSGTKSKHRGKIAMPHEYSWTISRTCLCIFIV CKEKAHRVIDRWMWMRGWKA
https://imgur.com/YVzQ2Q5
Information from this section can be freely copied and distributed. Everything for open science!