How Nature Solves Problems Through Computation

https://www.quantamagazine.org/how-nature-solves-problems-through-computation-20170706/

by digital55 • 9 years ago

243 73 9 years ago

komali22• 9 years ago

I absolutely love this way of thinking. We assume that an "individual" is a single human being, because it's convenient and I guess because that's how our sentience works. But realistically, the body itself is an extraordinarily complicated mass of "human" cells often at odds (see a cake, one part of the brain say "eat it," another says "dude no you'll get fat"), and thats without considering the masses of "non-human" entities, such as gut bacteria, skin bacteria, etc.

And then we can go macro - a tribe can subdivide into gatherers, warriors, and crafters. A city can specialize further. A country even further. Imagine how different the life experience would be if humans existed as single entities alone in endless fields.

Then it gets even more fun to consider interactions with other lifeforms - dogs, plants, cows.

Man, what a cool field of research. I'm glad to hear people are studying this.

fredley• 9 years ago

You may enjoy the game Everything. While the gameplay is straightforward, it's a very beautiful game, and is narrated by recordings of the philosopher Alan Watts.

didibus• 9 years ago

Totally. I've always been surprised that we don't consider how our bodies resemble much more a planet, then anything else. We're just cells to the earth, and who's to say mother earth isn't conscious. Our cells exhibit intelligence, they make their own decision and carry out actions independently, yet to our mutual harmonious benefit in general, but not always. Nature is much more then we make it be.

kr42• 9 years ago

Not just mother earth. Yogis have discovered in Samadhi (meditation trans) since ages that the universe itself is a giant body with us are like cells in it. Macrocosm is same as microcosm. http://omswami.com/2015/06/the-greatest-secret.html

This experience cannot be gain in a lab setting – you yourself are the lab, meditation is the tool and quieting the mind is the process to experience this.

lamlam• 9 years ago

and thats without considering the masses of "non-human" entities, such as gut bacteria

Actually, there's research that suggests that gut bacteria may have more control over our decision making then we are aware of. Really interesting stuff.

[1] http://www.smithsonianmag.com/science-nature/gut-bacteria-ma...

tomdre• 9 years ago

There's a nice TED book "Follow Your Gut: The Enormous Impact of Tiny Microbes" by Rob Knight

WhoBeI• 9 years ago

Ha! Yes. In 1.56 she convinced me their research is interesting, approachable and important. I also liked this from the projects page:

"Fundamentally, intelligence can be thought of as an organism's ability to sense, process, and respond to information to its adaptive advantage, forming memories to anticipate and optimize advantage in future encounters. By this definition, an amoeba has a rudimentary intelligence. So does a grove of aspen trees. Intelligence, then, isn’t the sole prerogative of humans, or primates, or even mammals, but a ubiquitous feature of life on earth."[0]

[0] https://www.santafe.edu/research/projects/evolution-of-compl...

naasking5• 9 years ago

Fundamentally, intelligence can be thought of as an organism's ability to sense, process, and respond to information to its adaptive advantage, forming memories to anticipate and optimize advantage in future encounters. By this definition, an amoeba has a rudimentary intelligence. So does a grove of aspen trees

But then what's an organism?

zitterbewegung• 9 years ago

What we call computers are merely artifacts that algorithms run on. A misconception that the name Computer Science pushes. But the artifact can be anything that follows the rules of an algorithm.

stareatgoats3• 9 years ago

"Fundamentally, intelligence can be thought of" - Personally I get grumpy when words get redefined in this way. The widely understood sense of the word (albeit a bit fuzzy) does at least not include the adaptive behavior of plants, and it's not like plants were just recently discovered to have such.

akvadrako• 9 years ago

Like most attempts to define intelligence, this one falls short. I hope we can agree that brute force doesn't count, since it's nearly a orthogonal trait. But physical strength contributes to an organism's "ability" to "respond to information".

will_pseudonym• 9 years ago

"We are a way for the universe to know itself." - Carl Sagan

throwaway91111• 9 years ago

I suppose, but this surprise seems grounded in the idea that a computation requires an individual. For instance, a physical description of computation might work if you defined it as a process that filters meaningful signals from entropic systems and expels the entropy from the system.

bane• 9 years ago

I think it's also worth thinking of humans as a kind of "menagerie" species. There are entire other species of animals who exist either on the input or output streams of human civilization: from cows to roaches.

It would have been much harder for humans to arrive at the current apex without these other animals.

From the inside to outside we're part of a larger whole.

pjmlp• 9 years ago

And some of them with some kind of intelligence we still do not fully understand (apes, dolphins, ravens, octopus, ...)

smallnamespace• 9 years ago

It would be great if we could get the point where we could look at the 'power spectrum' of a system and all its subsystems and make accurate predictions about how well it will perform in various scenarios.

For example, is it better to have one strong, centralized government, or a looser confederation of states? Is there an optimal amount of power one should delegate to individuals?

I mean, if you honestly compare the human body to society, our bodies look like horrible totalitarian states with a caste system attached.

Cells have their reproduction strictly controlled; cells that break those rules are induced to commit suicide (apoptosis) or are attacked and killed off by the police (immune system); cells that evade these measures and continue to reproduce often end up killing the individual (cancer). Even though cells (mostly) share all the same DNA, they are induced to specialize at birth and mostly stay in their roles for their entire lifespan.

So in what important ways are societies different from our bodies, such that societies should maximize liberty of the individual? What does that say about how we should structure society -- should it be a strict geographical hierarchy, or something looser (e.g. corporations and other groups), and how much power should go to each level of organization?

openfuture• 9 years ago

You could think of meditation as the dictator listening to what everyone has to say / letting them do their job. When the guy at the top doesn't listen then it makes sense for some group to make their own thing and grow to compete with the host government (cancer).

vbuwivbiu4• 9 years ago

There's been much work on this in graph and information-theory. To the point of centralization: it's a trade-off between efficiency and robustness (for one thing). Most living systems show evidence of this trade-off, having a mixture of centralization and distributed processing. For example, the body has a distributed blood supply (all cells share the same network) but has a central heart. Cells themselves have mixed centralized components (the nucleus, various organelles) and distributed aspects (the cyctoplasm, parallel transcription of DNA and production of proteins etc). The brain does distributed parallel processing but has centralized components (the amygdala, the pituitary gland, the cerebellum)

babyrainbow• 9 years ago

such that societies should maximize liberty of the individual?

This is required only because the controlling mechanism (the ruler/government) can be selfish and clueless in regards to what rules the subjects should follow for the well being of everyone. and also they are so detached from the problems of individuals....

Imagine a system where a ruler experience every pain and discomfort of every one of their subjects, real time (like in the body of a living thing). In such a system, I think individual liberty can be completely removed, because no one would not want to do anything outside of the rules anyway...

amelius• 9 years ago

I'm also wondering how systems of cells can produce feelings, and of what "strength" those feelings are bounded to be. Could it be that a system made from not cells but humans produce bigger feelings?

Anyway, regarding your comment, I'm not sure if we can say anything about how we should structure larger bodies (like societies) if we don't know anything about the feelings of each of their parts.

kr42• 9 years ago

Macrocosm is the same as microcosm. Universe is the living body with trillions of beings as its cells. Revealed countless times to yogis in Samadhi (meditation trans) since antiquity [0].

0: http://blacklotusapp.org/

youdontknowtho• 9 years ago

I'm sorry, I personally find a great deal of peace in Buddhist writings, but let's not start taking artistic philosophical language and equating it with scientific observation.

The microcosm and the macrocosm are vastly different with one exhibiting quantum behaviors and the other mostly influenced by gravity.

The experience of yogi's or Buddhists of a removal of psychological barriers that separate subjective experience into a personal and extra-personal world are just that. Changes to subjective experience. That means a change in brain state. It doesn't imply some kind of magic that changes the observed world.

Because things seem alike in imagery or language doesn't even remotely mean they are alike in observation.

What you are saying isn't unlike someone equating the Bohr model of the atom with a solar system. They do look alike in illustrations, after all. The problem is that solar systems don't look like atoms and they are vastly different "things".

cel1ne• 9 years ago

We assume that an "individual" is a single human being, because it's convenient and I guess because that's how our sentience works.

I often hear that, but I think that's a wrong assumption. I think people fall in and out of being conscious all the time, and often they are semi-conscious in a distracted way.

Consciousness is this holy singular grail, but I think it's actually a muddy mixture of state of minds, like thinking usually is.

vbuwivbiu4• 9 years ago

and, as Scott Aaronson points out, why assume the brain only hosts a single consciousness* ? Maybe there are multiple consciousnesses hosted by each person's brain, and "you" are one of them at any given time - you'd have no idea the others are conscious.

* (not to be confused with multiple-personality disorder)

unabst• 9 years ago

"Individual" and "single human being" are abstractions. So are "cells" and "skin" and everything else. If you remove abstraction, you remove language, but an undivided landscape of activity of nature remains. Language is just a layer and for our convenience.

But where ever there is communication, abstractions are required to compose a coherent and persistent message. So DNA and proteins and even subatomic particles that entangle can be considered to be tapping into natures capacity to abstract, and to compute those abstractions.

And if you consider nothing exists in a void, and the innate connected-ness of an undivided and unabstracted universe as a whole, everything is coming and going; everything is itself a message and the subject of computation; and nature is engaged in a never ending conversation with none other than itself. If it's messages can be any size, and last however long, then you and I too are just one of its many messages.

(pardon the philosophy... stuck doing mundane work at midnight on a Friday... I had to write something)

ldd• 9 years ago

Look into pantheism if you ever have the time.

enkid• 9 years ago

It's abstractions all the way down. There is no inherent concept, including atoms, photons, matter, or energy. It's just how an intellectual system can process what it experiences.

21• 9 years ago

Quote from an article along the lines:

However, in broad brush, we might say this: You’re a pattern in spacetime. A mathematical pattern. Specifically, you’re a braid in spacetime—indeed, one of the most elaborate braids known.

http://nautil.us/issue/9/time/life-is-a-braid-in-spacetime

https://news.ycombinator.com/item?id=10323222

nikki934• 9 years ago

Consider the dual nature of CSP and Actor models and how it reflects wave particle duality of information packet movement in a field.

j2kun2• 9 years ago

The last few years have seen a number of very interesting developments along the lines of understanding the natural world through the so-called "lens of computation." Some interesting talks can be found here [1] as well as an essay here [2], on the topics of economics, social interaction, biology, and physics.

[1] https://www.ias.edu/ideas/2014/lens-of-computation-workshop

[2] http://theory.cs.berkeley.edu/computational-lens.html

beefman• 9 years ago

Generally that phrase refers to the asymptotic analysis of algorithms, which I think has relatively little to teach us about natural systems of the kind mentioned here. I have tremendous respect for TCS and those who study it, and I attended the second "lens" conference in Berkeley in 2011. But it does seem a bit like a field looking for justification. And I don't think it's impossible that academic competition drives more of our brightest to spend more time on it (asymptotic analysis) than is optimal.

j2kun2• 9 years ago

Why would you think it has little to teach? The author even says that it's useful and influential.

What we are doing at C4 is taking messy, conceptually challenging problems and turning them into something rigorous. We’re very philosophically oriented, but we’re also very quantitative, particularly in thinking about how nature can overcome subjectivity in information processing through collective computation. We really think the answer to these questions requires combining insights from statistical physics, theoretical computer science, information theory, evolutionary biology and cognitive science.

jchanimal• 9 years ago

they got data by inducing monkey fights.

""" We were interested in whether we could induce the monkey society we were studying to change from its status quo of many small fights and a few large ones to having many large fights. We observed that fights in this monkey group range in size from two to 30 or so individuals, with small fights common and large fights very rare. By simulating the society using data we had collected on fight-joining decisions, we found that we could measure the number of monkeys whose propensity to join fights would have to increase to move the system closer to the critical point.

openasocket• 9 years ago

That sounds like they were just simulating the society based on data they had gathered about their fights. So not inducing monkey fights, just observing the monkeys fighting naturally and extrapolating from there. Though earlier in the article they allude to removing a few members of the society and showing the fights increased.

Also it should be noted that fighting is a very common part of monkey behavior. If you observe these groups in the wild individuals are almost constantly challenging and checking others to maintain or advance their position in the social hierarchy.

mlechha• 9 years ago

It was a simulation from the data they already had. It's near impossible to do that with a real monkey colony.

debbiedowner2• 9 years ago

Why is a "social coordinate space" not Euclidean, and why is it worth mentioning?

..."phenomenological rules"...

It is an interesting subject... but this interview is a little annoying since it is with a philosopher in scientists clothing.

https://www.linkedin.com/in/jessica-c-flack-76b454133/

dasimon• 9 years ago

A prior paper contributed to by Ms. Flack (https://arxiv.org/ftp/arxiv/papers/1406/1406.7720.pdf) presents a framework for modeling the dynamics between individuals - I would assume that is the abstract "social coordinate space" to which she's referring, which places individuals within "Markovian, probabilistic, 'social' circuits" rather than into some Euclidean space.

And Ms. Flack, though her research interests extend beyond the realm of what is traditionally considered hard science, is most certainly a scientist - an evolutionary biologist specifically.

I assume you are linking to her LinkedIn profile to point out her "Doctor of Philosophy" degree? That's just the full, formal name for a PhD :P

debbiedowner2• 9 years ago

Just to get the full quote from original article: “their metric space is a social coordinate space. It’s not Euclidean.”

If the space being considered is a probability space like you say, then there is no metric, there is a measure.

---

Why am I so pedantic?

Because I feel that the liberal arts mindset of writing is not serious enough for subjects where there is a ground truth to uncover.

In the article you link, where is the empirical study that compares how this model predicts reality?

I wish that people who publish on SIAMS, or ASA, AMS, IEEE affiliated journals were the ones trending on HN, and getting interviewed by journalists. A pessimist would say those people are too busy with their work for publicity, and that an empty barrel echoes loudest.

ehudla3• 9 years ago

While this field is advancing rapidly, and sophisticated approaches are being constantly developed, for those who want a gentle introduction, I venture to recommend our book: https://www.amazon.com/Biological-Computation-Chapman-Mathem...

Section 6.1 deals with swarm intelligence (ant colony optimization, cemetery organization, particle swarm optimization). 6.2 is on artificial immune systems.

A more biological+philosophical exploration on issues of collectivity will be found in the following book, which is already available for pre-order on Amazon, I see:

https://www.amazon.com/Landscapes-Collectivity-Sciences-Theo...

shishy3• 9 years ago

Are there any academic research papers that you would recommend in this field?

ehudla3• 9 years ago

Too many... What topics specifically interest you? I might have suggestions, if it's topics I am familiar with.

known2• 9 years ago

"But when Flack removed some of the police, the whole group became fractured and chaotic"

Reminds me Saddam and Iraq :)

known2• 9 years ago

I believe UNSC does the same;

fithisux• 9 years ago

Can someone suggest FOSS code for experimentation? Any free tutorials, notes, books.

Thanks

How Nature Solves Problems Through Computation

Related Stories