What's the difference? If both systems can be modelled in a similar way, how does that not make them similar?
I don't think anyone is trying to make an argument that they are the same 'thing' ("hey, maybe the observable Universe is a 'thought' inside a cosmic brain!"), just that there are striking similarities in structure - which can only easily be seen by a comparison of models.
The differences are major. One is made of unconnected stars, largely governed by the laws of gravity, the other one is made of connected cells, governed by the laws of chemistry, electricity (and maybe quantum effects).
The only rough similarity is the looks, and only if you dye the brain tissue in a special way.
I just think that is why it is so interesting; that they share a similar structure even though they are formed in completely different ways may not tell us anything about the Universe but it is still something to note.
I have a little hypothesis that I dreamed up a couple nights ago:
Life is exogenous to the universe. That gives life it's meaning-to define what it is through it's impact on the universe. This is because the universe is in average empty and lifeless. The inanimate universe vastly outweighs the animate. Life could not ever overcome the average emptiness of the universe. But still it must try. It must try because if it ceases it becomes the inanimate. Life is that exogenous influence on the universe not attributable to any other physical or chemical process. If you simply cannot explain it then it may be life.
Brain evolution over time is definitely dependent upon physics/chemistry - but it's also dependent on conditions at higher levels of abstraction. We know neurons change their associations every day based on many factors at the conscious and subconscious levels (observations/thoughts).
>What's the difference? If both systems can be modelled in a similar way, how does that not make them similar?
I think that this is a good question, and central to a lot of the confusion that stems from a casual reading of some modern topics in physics. It's the difference between a black hole in nature, which formed in a realistic process, and the Schwarzschild metric, which has a uniform sphere of "dust" collapsing from infinity, in an empty universe, without gravity, and no angular momentum or charge.
The reason these models are used is because they can give insights and/or generate new tools to explore reality with. On the other hand, they're just maps, just approximations of systems which we couldn't hope to accurately model with fidelity.
On one hand, you get some amazing insights when your models intersect with observations, but you also need to remember... "The map is not the territory."
I'm somewhat confused. Are you saying that there is no relation between the structure of the large-scale Universe and the structure of Neurons? If so, then why do we get two models that come out looking very similar?
Of course, it is unlikely that there is a physical, causal link between the two. But if I have two photographs of two different things and they look similar, is it not then fair to say that the way those two different things look is similar?
[Edit: also - to be fair - it is not just models of structure formation (like the power law presented in the article) - we have pretty pictures of the large scale structure of the Universe as well http://www.sdss3.org/images/gallery/sdss_pie2.jpg]
To use the photograph analogy, if a person uses a typical 35mm camera a photo of a duck will have the same proportions and color gamut as a photo of an iPhone and have a similar tonal range and if the focal length of the lens is the same, the field of view will be the same.
The model is an instrument not the subject itself. A person cannot cook dinner in the kitchen of a blueprint.
But that is not how these models are constructed. They are from empirical data - they just tell you how many objects of a given size exist. The analogy is that the power spectra for the brain and the power spectra for the large scale structure are two different photographs (of actual things!) and look strikingly similar.
Viewed from five meters away, a photo of a duck and a photo of an iPhone may appear to be strikingly similar. The important differences depend on the level of granularity at which they are observed and the purpose for which they are observed.
All models are wrong but some are useful. -- George Box
The article does not demonstrate the utility of the (non-wrongness) similarities between the models. It address the "A cow is like a hammer because there is a 'b' in 'both'," possibility. Two models developed using similar techniques used in the discipline of modeling will have non-functional similarities.
I don't think that anyone is trying to argue that the similarity between the power spectra is useful but that doesn't mean it isn't interesting. Also, the level of similarity between them is striking!
To me, it's basically p-hacking. Because the it starts with human artifacts the similarity is more on the level of finding that two English language texts have similar letter frequencies and presenting it as an episode of Ancient Aliens.
Eh, I disagree. Power spectra are a pretty standard thing to use to compare e.g. different structure formation models in Cosmology and I still stand by the idea that the two are strikingly similar (and how interesting that is).
> Programs like the Human Brain Project, designed to simulate an entire human neuronal network
I'm not sure but it is my understanding the HBP plans on studying the human brain from a static measurements of adult specimens. That seems at odd with the cosmological models used here, which are dynamical, building a structure from an initial conditions.
Is there an accurate model for the dynamics of brain during the development of the embryo? If not couldn't we start from the cosmological model and tweak it until it produces a structure that matches the brain even better?
The argument here is a bit forced. In fact, the entire universe seems to be self-similar. My point is only that there's nothing special, mysterious, or spooky about brain matter (as the article kind of implies). In fact, you can draw similarities between atoms (and their electron clouds) and solar systems.
The self-similarity and fractal nature of our world is indeed puzzling and fascinating, but it goes far beyond brains.
It's not that profound. Both networks form through "preferential attachment." Large networks of galaxies form because gravity causes the largest conglomerations to get larger. Similarly, neurons are more likely to connect to other neurons that already have a large number of connections. It's the same principle exhibited by different mechanisms; also the reason social networks look the way they do. See Barabasi-Albert model, preferential attachment, and scale-free networks.
My thought as well. A bunch of nodes that aren't all connected has got to have some sort of selection mechanism. Whether that's gravity or ions or social status, it will end up looking like that, a few nodes that are highly connected and many that are not.
If you read some complexity books this is the theme: the same structures show up in different substrates.
Rather than "profoundness" I'd argue that there's something inherently beautiful about it. Granted, beauty is not a objective measure of anything in particular, and "we are like starstuff" might be an overly done cliche, but we humans are mostly driven by subjective and aesthetic notions, especially when we're young. If we taught more children that we are -really- like stars and how, with pictures and math they can grasp, we'd inspire more of them to get into science and technology, perhaps even bring spacefaring tech and the stars themselves a little closer to us in the long run.
I personally treat profundity as synonymous with a large first derivative of compressibility (see Schmidhuber). Something is a profound insight if it allows to compress information much more than before, e.g. by explaining many phenomena with a simple rule (see physics). In this case, there is not much to compress. The two phenomena have closely resembling spatial features, but one is explained by the gravitational force and the other one is likely explained by evolutionary necessity. Maybe one can find some superficial relations via scale-free networks etc., but my expectations that there is anything to gain from that are very low, just because of how different the underlying mechanisms are (matter moving by attractive forces vs biological stuff that needs to communicate efficiently such that its superstructure does not die until it procreates or until its descendants are sufficiently autonomous for self-replication or to execute the adaptations shaped by some symbiotic, altruistic or group-selective evolutionary correction signals, if any).
I think what's beautiful is that it shows the endless fractal self similarity of our reality in general. You see it absolutely everywhere in everything. The curl of Maxwell's equations gives us both the behavior of a single photon in your eyeball, to the distribution of trillions of tons of plasma in a Quasar. We truly are "all one" in a very literal sense.
Yet nature is not scale-invariant. E.g., you can't take a carbon atom and scale it by a factor of 10 and make it behave in the same way as the original atom.
>"Yet nature is not scale-invariant. E.g., you can't take a carbon atom and scale it by a factor of 10 and make it behave in the same way as the original atom."
That's a major caveat, although I think in the end it just makes things more interesting. "There's plenty of room at the bottom" as Feynman said. Technologies like nanophotonics will allow us to have things like sold state LIDAR and perfect spatial tracking for virtual reality. Also to boost solar cell efficiency to ranges >50%, which has been impossible with the classical light modeling used up until now.
Perhaps the reality is that all of these scales are linked fundamentally, but proving that would be quite a task.
This article seems quite silly, or at least, can easily lead to silly conclusions. There was a brief paragraph at the end that half-heartedly admitted that the evidence wasn't conclusive, but it seems most people didn't read it.
Further, only one paper was referenced - and it was written by the author of the article! Guess what? It has 1 (one) citation.
>Is the apparent similarity just the human tendency to perceive meaningful patterns in random data (apophenia)? Remarkably enough, the answer seems to be no: Statistical analysis shows these systems do indeed present quantitative similarities.
This is not a good argument. The similarity of images is not the similarity of systems.
Also,
>In other words, it tells us how many high-frequency and low-frequency notes make the peculiar spatial melody of each image.
.. come on, man. Come on.
>Based on the latest analysis of the connectivity of the brain network, independent studies have concluded that the total memory capacity of the adult human brain should be around 2.5 petabytes, not far from the 1-10 petabyte range estimated for the cosmic web!
So if I make a 3-petabyte hard drive, it's self-aware? There are a few things wrong with this passage. The first is that no citations are given. The second is that connectivity is not the only important aspect of brain function. The third is that 2.5 petabytes falls firmly WITHIN the 1-10 pb range given. The fourth is that a 1-10 pb range hardly seems to tell us anything.
>Roughly speaking, this similarity in memory capacity means that the entire body of information that is stored in a human brain (for instance, the entire life experience of a person) can also be encoded into the distribution of galaxies in our universe.
No. Galaxies don't fire action potentials. The data points could be mapped to each other, perhaps, but this doesn't mean that the information is the same, given that the information depends on the way the data is read.
>It is truly a remarkable fact that the cosmic web is more similar to the human brain than it is to the interior of a galaxy; or that the neuronal network is more similar to the cosmic web than it is to the interior of a neuronal body
Is it really? Seems like apophenia is taking hold again. Of course we want to recognise ourselves in the universe. God made man in his image blah blah blah blah. People should be wary of this way of thinking.
43 comments
[ 3.0 ms ] story [ 40.9 ms ] threadThis is a similarity between models of neuron and galactic networks, not the actual networks themselves.
I don't think anyone is trying to make an argument that they are the same 'thing' ("hey, maybe the observable Universe is a 'thought' inside a cosmic brain!"), just that there are striking similarities in structure - which can only easily be seen by a comparison of models.
The only rough similarity is the looks, and only if you dye the brain tissue in a special way.
Life is exogenous to the universe. That gives life it's meaning-to define what it is through it's impact on the universe. This is because the universe is in average empty and lifeless. The inanimate universe vastly outweighs the animate. Life could not ever overcome the average emptiness of the universe. But still it must try. It must try because if it ceases it becomes the inanimate. Life is that exogenous influence on the universe not attributable to any other physical or chemical process. If you simply cannot explain it then it may be life.
Brain evolution over time is definitely dependent upon physics/chemistry - but it's also dependent on conditions at higher levels of abstraction. We know neurons change their associations every day based on many factors at the conscious and subconscious levels (observations/thoughts).
I think that this is a good question, and central to a lot of the confusion that stems from a casual reading of some modern topics in physics. It's the difference between a black hole in nature, which formed in a realistic process, and the Schwarzschild metric, which has a uniform sphere of "dust" collapsing from infinity, in an empty universe, without gravity, and no angular momentum or charge.
The reason these models are used is because they can give insights and/or generate new tools to explore reality with. On the other hand, they're just maps, just approximations of systems which we couldn't hope to accurately model with fidelity.
On one hand, you get some amazing insights when your models intersect with observations, but you also need to remember... "The map is not the territory."
Of course, it is unlikely that there is a physical, causal link between the two. But if I have two photographs of two different things and they look similar, is it not then fair to say that the way those two different things look is similar?
[Edit: also - to be fair - it is not just models of structure formation (like the power law presented in the article) - we have pretty pictures of the large scale structure of the Universe as well http://www.sdss3.org/images/gallery/sdss_pie2.jpg]
The model is an instrument not the subject itself. A person cannot cook dinner in the kitchen of a blueprint.
The article does not demonstrate the utility of the (non-wrongness) similarities between the models. It address the "A cow is like a hammer because there is a 'b' in 'both'," possibility. Two models developed using similar techniques used in the discipline of modeling will have non-functional similarities.
http://freakonometrics.hypotheses.org/19817
http://files.abovetopsecret.com/images/member/016c28958aab.j...
http://wwwmpa.mpa-garching.mpg.de/galform/millennium/poster_...
Does that mean our solar system is Oxygen or Fluorine? Or are gas giants neutrons?
Are galaxies cells? Do the orders of magnitude map?
Are we then still the universe experiencing itself?
I'm not sure but it is my understanding the HBP plans on studying the human brain from a static measurements of adult specimens. That seems at odd with the cosmological models used here, which are dynamical, building a structure from an initial conditions.
Is there an accurate model for the dynamics of brain during the development of the embryo? If not couldn't we start from the cosmological model and tweak it until it produces a structure that matches the brain even better?
The self-similarity and fractal nature of our world is indeed puzzling and fascinating, but it goes far beyond brains.
https://en.wikipedia.org/wiki/Macrocosm_and_microcosm
If you read some complexity books this is the theme: the same structures show up in different substrates.
Also, this validates cosmic brain memes.
how do neurons "know" which neurons have large number of connections?
That's a major caveat, although I think in the end it just makes things more interesting. "There's plenty of room at the bottom" as Feynman said. Technologies like nanophotonics will allow us to have things like sold state LIDAR and perfect spatial tracking for virtual reality. Also to boost solar cell efficiency to ranges >50%, which has been impossible with the classical light modeling used up until now.
Perhaps the reality is that all of these scales are linked fundamentally, but proving that would be quite a task.
https://en.wikipedia.org/wiki/Power_law
This discussion, yesterday, on normal distributions, also has some very good insights..
https://news.ycombinator.com/item?id=14816130
surface of the sun: https://cdn.images.express.co.uk/img/dynamic/1/590x/thesun-4...
Further, only one paper was referenced - and it was written by the author of the article! Guess what? It has 1 (one) citation.
>Is the apparent similarity just the human tendency to perceive meaningful patterns in random data (apophenia)? Remarkably enough, the answer seems to be no: Statistical analysis shows these systems do indeed present quantitative similarities.
This is not a good argument. The similarity of images is not the similarity of systems.
Also,
>In other words, it tells us how many high-frequency and low-frequency notes make the peculiar spatial melody of each image.
.. come on, man. Come on.
>Based on the latest analysis of the connectivity of the brain network, independent studies have concluded that the total memory capacity of the adult human brain should be around 2.5 petabytes, not far from the 1-10 petabyte range estimated for the cosmic web!
So if I make a 3-petabyte hard drive, it's self-aware? There are a few things wrong with this passage. The first is that no citations are given. The second is that connectivity is not the only important aspect of brain function. The third is that 2.5 petabytes falls firmly WITHIN the 1-10 pb range given. The fourth is that a 1-10 pb range hardly seems to tell us anything.
>Roughly speaking, this similarity in memory capacity means that the entire body of information that is stored in a human brain (for instance, the entire life experience of a person) can also be encoded into the distribution of galaxies in our universe.
No. Galaxies don't fire action potentials. The data points could be mapped to each other, perhaps, but this doesn't mean that the information is the same, given that the information depends on the way the data is read.
>It is truly a remarkable fact that the cosmic web is more similar to the human brain than it is to the interior of a galaxy; or that the neuronal network is more similar to the cosmic web than it is to the interior of a neuronal body
Is it really? Seems like apophenia is taking hold again. Of course we want to recognise ourselves in the universe. God made man in his image blah blah blah blah. People should be wary of this way of thinking.