Playing With Particle Life

[00:00:00]Just recently, I've been playing around with a thing called particle life. This video is about what it is and what we can perhaps learn from it.

[00:00:12]Just a quick heads-up, some of the scenes in this video include patterns that may be triggers for migraine, trypophobia, or other visually stimulated conditions. If you're sensitive to those sorts of things, this video might not be for you. As is often the case with these sorts of things, this thing has existed for several years and I'm very late to the party, but here we go. This is particle life. It's a website where, as you can see, there are a load of little colored dots swarming about, interacting with each other, forming interesting larger shapes, and those larger shapes sometimes move or swim about as though they were living things.

[00:00:47]The stuff that's going on here may appear to be quite complex, and it is complex. In fact, you might look at these things swimming around and think maybe someone specifically programmed them to do this, to look and behave the way they're doing.

[00:01:00]But, these things decided to do this all on their own. Everything you see here is all based on a particle simulation with a quite small number of relatively simple rules.

[00:01:11]It's really a highly simplified physics simulation, and it's an interesting demonstration of how complex behaviors can emerge from simple building blocks.

[00:01:20]So, it's a simulation where a collection of particles are all interacting, but why is it called particle life? Are these things in here actually alive?

[00:01:29]Well, no. Or at least, probably not. I say that because it turns out that clearly and unambiguously defining what we mean by alive is harder than it might seem. It's actually very difficult to come up with a rigorous, complete definition of living things that doesn't either exclude certain things that we intuitively know to be alive or accidentally include things that we don't generally consider to be living things. For example, a common definition of living things might be that they are self-contained or bounded. That is, there's a place where they are and a place where they're not. There's a boundary between them and the rest of the universe. In the case of bacteria, that's the cell membrane. In the case of humans, it's the skin and so on. They consume things or process energy. They have a metabolism that enables them to continue to actively exist by exploiting chemical reactions. So, ingestion and digestion. And reproduction. That is, they can make copies of themselves. The trouble with this definition is that it excludes things like individual worker ants, which are not capable of reproduction, even though nobody would dispute that they're alive.

[00:02:36]And this definition includes things like fire. A flame has boundaries, but it exploits chemical reactions to sustain itself, and it can reproduce. A fire can shed sparks that will start new fires.

[00:02:48]And yet, nobody thinks that fire is really alive.

[00:02:52]And it doesn't seem to matter how you tweak the terms of the definition, you always seem to end up including something you don't want to include or excluding something you do want. Alive or living things, it's one of those things where really we know it when we see it. Anyway, that's a bit of a side track because this thing is called particle life, not because anyone wants to believe that these little creature things are actually alive, but because it draws inspiration from prior art.

[00:03:18]Specifically, John Conway's Game of Life. Sometimes just referred to as Conway's Life or just Life. So, let's take a really quick look at what that is.

[00:03:27]John Conway was a very talented English mathematician who contributed a great deal of interesting insight to a wide range of mathematical fields, as well as theoretical physics.

[00:03:36]But he is most widely known as the inventor in 1970 of a cellular automaton called the Game of Life, which became a hugely popular thing, and indeed Conway grew to rather dislike the fact that this thing was the thing he was best known for, and felt that the attention it received overshadowed many of his other achievements. I'll link to the Wikipedia page about him where you can start reading about some of those other things. It's quite the rabbit hole.

[00:04:00]Anyway, Conway's life works like this.

[00:04:03]There is a grid of squares or cells.

[00:04:05]This is the world. And a cell can either be alive, that is filled in, or it can be dead, that is empty. What happens to any cell is determined by the state of its eight neighbors.

[00:04:16]If a living or filled cell has two or three neighbors, it remains unchanged.

[00:04:20]It survives.

[00:04:21]If it has less than two neighbors, it dies of isolation.

[00:04:25]If it has four or more neighbors, it dies of overcrowding.

[00:04:29]And if an empty cell is surrounded by exactly three neighbors, a new cell is born there.

[00:04:34]And that's it. Just a few really simple rules. And yet when these rules are applied repeatedly to a grid containing an initial pattern of cells, you get this sort of thing.

[00:04:43]Apparently quite complex things happening, but all based on that handful of simple rules.

[00:04:49]As we can see in this example, a really simple starting pattern grows and develops into what in some ways resembles a colony of living things like bacteria or something.

[00:04:57]And include some things that appear to function like discrete organisms. Here's one of the simpler and better known examples of that. This thing's called a glider, and it's a pattern that cycles through several states walking across the grid as it does.

[00:05:11]Now you're watching an implementation of Conway's life in a web browser here, but there have been many implementations of this over the years. In fact, the popularity of this thing is partly driven by the fact that this is a really easy thing to implement on almost any computer.

[00:05:25]But it is interesting to note that when Conway first designed it, it was not a computer program. He was doing the computation manually in pencil on sheets of squared paper. A computer is optional for this.

[00:05:36]So that's Conway's life, and particle life is called particle life because it takes inspiration from it.

[00:05:42]But particle life isn't a cellular automaton. It doesn't take place in a grid of discrete cells. It's a simulation consisting of a collection of particles of assorted colors. Each color having its own defined ways of interacting with all the other colors of particles. Let's look at how that would work with only two particles of two different types, red and green. Each particle type has a few parameters with respect to the other type. Firstly, whether it's attracted to the other particle or repelled by it, and also the minimum and maximum range over which that attraction or repulsion is effective.

[00:06:13]The range parameters are somewhat of an engineering necessity just so the simulation doesn't have to calculate the interaction with every single other particle in the world, even if they're far away. But also within their ranges, those forces have gradients, just like real-world magnetic and electrostatic forces do.

[00:06:29]Nearer things are pushed or pulled harder than further things. So, if a particle can only attract or repel, what stops them from either all clumping together or all flying apart? The answer is the defined behaviors of the other types of particles. So, for example, we might say that red is attracted to green at a relatively wide range, but we can also define that green is more strongly repelled by red at a close range.

[00:06:52]The result of this is that red will try to get close to green, but when it does get close, green will try to get away.

[00:06:58]Red will continue to try to get close to green and will end up chasing green across the screen.

[00:07:04]That's just the simplest possible illustration with only two particle types and only one of each type of particle. If you've got multiples of each particle type, then you need another parameter to describe how green reacts to green and how red reacts to other reds.

[00:07:17]So, with those additional parameters, you not only have movement emerging, but also structure. For example, if red is attracted to green and itself, it will clump up and go chasing green.

[00:07:27]If green is repelled by red and itself, it will tend to space itself out, but when it's captured by red, it will still space itself out in a shell around a core of red. The initial positions of all of the particles are random, so the arrangement of these particles doesn't typically happen perfectly symmetrically. And then, instead of neat clumps, you get oscillating blobs or simple things that look like swimming or swirling creatures.

[00:07:51]But, you're not limited to two types of particle. And for each particle type, there is a set of interaction parameters for all of the other types.

[00:07:59]This is managed in a matrix that looks like this. And you can tweak all of those parameters to try to make things behave in specific ways.

[00:08:05]But, often it's just as much fun to randomize everything and watch how it develops based on these randomly assigned parameters.

[00:08:12]If most of what you have is repulsive forces, things tend to spread out into a static pattern. Sometimes quite an interesting one.

[00:08:19]If most of what you have is attractive forces, things tend to clump up into a sparse set of little balls.

[00:08:26]It's when you have a more even mix of attraction and repulsion that the most interesting stuff tends to happen.

[00:08:31]Because you've got particles that really want to be next to other particles that really don't want that, it causes activity.

[00:08:39]There's also friction in the simulation in general, so things tend to damp down.

[00:08:43]The alternative would be prone to really noisy and chaotic agitation.

[00:08:48]Most often, you start with a random scattering of different particles and it starts to develop with little swarms of particles moving around, gathering up more of their like kind into clusters which form stacked layers chasing each other.

[00:09:00]It very quickly begins to resemble a world full of squishy biological swimming creatures zooming about, eating things and eating or avoiding each other.

[00:09:10]Often, it reaches a point where some of these creatures have collected up so many particles that their gelatinous wobbling becomes chaotic and they stop swimming and pulsate violently and then explode, shedding particles that arrange themselves into smaller and simpler structures which may combine to form new creatures that swim off on their own. I've called this movement swimming, and it is very tempting looking at this to think of these particles or entities as being suspended in some sort of clear fluid.

[00:09:37]The way things move around really does suggest that.

[00:09:40]But there is no substrate. There's no liquid here. The fluid is the simulation itself. The particles are the fluid.

[00:09:47]And when they join together, they are the objects in the fluid, too.

[00:09:50]So it does end up being a fluid simulation kind of by accident just by virtue of being a physics and I suppose a sort of chemistry simulation.

[00:09:58]Very often, one example of one arrangement of particles will dominate.

[00:10:03]Rapidly moving across the screen, consuming or combining with groups of particles that it itself contains. And when this happens, it can agitate the rest of the world so much that all the other things don't get much of a chance to develop before that big thing comes galloping through and smashes everything apart.

[00:10:19]It really is quite interesting to watch.

[00:10:21]And the illusion that these are little life forms swimming in a pond or something is really rather compelling.

[00:10:27]Sometimes a set of randomly assigned rules won't actually create discrete roaming creatures, but instead an undulating pattern of waves or a field of cell-like entities in a more constrained structure or things that look like chains or vessels or such.

[00:10:43]But if we zoom in here, we can see it's all just a load of oscillating feedback loops of attraction and repulsion.

[00:10:49]Sometimes the behavior of these collection of particles is driven by really subtle little things. For example, in this one, there are only two types of particle and they are clumping together. Then the clump moves in a straight line for a short finite distance. Then the whole thing suddenly splits in two and the two pieces move in opposite directions until they combine with another clump. Then they repeat the move and split thing. It all looks very orderly and defined.

[00:11:14]But what's curious here is the consistency of the distance they move before suddenly dividing. It's almost like there's some timed event happening.

[00:11:21]But there is no explicit timer property in the individual particle behaviors.

[00:11:26]Instead, I think what's happening here is that the clumps of particle accelerate. And as they do so, they change shape, flattening out a little, and this places some of the particles far enough away from each other that they can escape an attraction force that maybe has a shorter range, and then as they pull away, they rip the clump in two. But, this can only happen after the clump changes in shape a little bit as it moves.

[00:11:48]Now, there's nothing in here that's happening without a complete explanation. All of these effects have distinct, explicit causes.

[00:11:55]And everything here is just behaving completely mechanically according to whatever parameters are set.

[00:12:00]The simulation can run in two or three dimensions. The 3D version is interesting, but I find the 2D world easier on the eyes and easier to understand more of what's going on.

[00:12:11]So, what's the point of all of this?

[00:12:13]Well, if you want, there doesn't need to be a point at all. It can just be an interesting toy or an appealing bit of animated wallpaper. But, if you're interested, this is also a way to grasp or play with emergent phenomena.

[00:12:24]That is, how apparently complex and organized large-scale things can arise as the result of a very large number of very small things, all individually following a simple set of rules.

[00:12:36]Similar sorts of things appear in the way birds flock together. Here's a murmuration of starlings, the way insects swarm, and the way fish form shoals.

[00:12:45]It may look like an organized, intelligent, large entity, but it's all happening because the parts are following simple, local rules.

[00:12:53]But, on a much smaller level than that, the way atoms and molecules interact with one another to make it possible for living organisms to exist is like this.

[00:13:01]All of the atoms in your body are just doing what atoms are supposed to do.

[00:13:05]They're following a set of simple, predictable behaviors in the way they interact with other atoms in your body, and in the air you breathe, and the food you eat.

[00:13:13]But, the sum of all of those interactions is an unfathomably complex and nuanced emergent phenomenon. It's not random. It's, in fact, very organized and orderly.

[00:13:23]But, it's based on simple rules all the way down to atoms, and indeed, all the way further down to the things that make up those atoms.

[00:13:30]There are limits to the complexity of the stuff that can arise in particle life just because it's a tiny microcosm.

[00:13:36]There are only 20 types of particle, only one force, no valencies or ionic bonds or atomic weights, and the maximum number of particles at play is limited to a few hundreds of thousands. In practice, maybe fewer than that unless you have a very powerful computer.

[00:13:52]That's many, many times fewer than the number of atoms in the smallest speck of dust you could possibly see.

[00:13:58]So, there are lots of reasons why we shouldn't expect to see actual analogs of living organisms arise here, or even analogs of normal chemistry. And there's not even space to create those things on purpose.

[00:14:08]But it is, I think, still a useful demonstration of how complex things can be built, or rather, how they organize themselves from simple pieces just following simple, straightforward rules.

[00:14:20]I've mentioned before on this channel that I don't play a lot of actual video games, but something like this, just a simple sandbox, not a game with any specific objective, more like a sort of construction toy or desk ornament, can keep me occupied and fascinated for hours.

[00:14:35]The rest of this video will just be a gallery of some patterns that I recorded that I found pleasing to watch, together with some ambient music.

[00:14:42]Thanks for watching, and I hope to see you again soon.

[00:15:52]>> [music] [music] [music] >> Yeah.

[00:17:48]>> [music] >> Hallelujah.

[00:18:07]Hallelujah.

[00:19:16]>> [music] [music] [music] [music] >> Mhm.