Hello, I'm Stephen Wolfram and I'm going to talk about Biological Evolution and other Adaptive Processes.
Just kidding, you know I'm here to talk about cellular automata.
breck 4 days ago [-]
This is one of my favorite SW essays. A tour de force in terms of using CA to show evolution, mutation, and natural selection.
Don't miss the "Personal Notes" section at the bottom. Neat history and insight into the process and life of a scientist.
> In a biological analogy we might have a genome that seems to lead to unbounded perhaps-tumor-like growth—but where actually the growth in the end “unexpectedly” stops.
I liked this aside. Early "cancer" detection can often lead to harm, mistaking benign for malicious growths.
> somehow follow from natural selection a general law...that would lead to the evolution of progressively more complex organisms, culminating in the “higher animals”. But absent the kind of model I’m discussing here, nothing in the later development of traditional evolutionary theory really successfully supported this—or was able to give much analysis of it.
A thing that comes to mind is Assembly Theory by Leroy Cronin and Sara Walker.
andoando 4 days ago [-]
I love this stuff but Im not understanding the diagrams here at all from the very beginning.
rtkwe 4 days ago [-]
It's a cellular automata where each line is 'offspring' of the line above it. Breaking down the very first image [0]; on the left we have the rules that determine what each cell on the next line should be based on the line directly above it. Taking the first one for example if the box above, above and right and above and left are purple the cell is purple. To the right of that we see a full run of that automata starting from 1 orange box to all white boxes after which that set of rules is in a sense dead because it will just be white forever because of the final rule.
From there it's taking this and applying some pretty old genetic programming concepts to it to illustrate how we can see evolution in this every simple system. Wolfram has been very interested in cellular automata for a long time trying to find interesting ways to use it or study it even writing a gigantic book about it [1].
[1] Called a new kind of science, he's made it available online if you're interested in learning about automata in maybe the deepest depth with some fun but ultimately not very useful theoretical applications: https://www.wolframscience.com/nks/
tedmw 4 days ago [-]
Yeah, it’s not explained very well. If you click the “Cellular automata” link [1], that paper explains it in more detail. A cell’s next state is determined by its color and its neighbors’ colors.
For example, the top left corner of the key shows that if a cell is blue and its neighbors are blue, it will remain blue. The one to the right shows that if a cell is blue and has a blue cell on the left and a red cell on the right, it will switch to white.
Ah ok, thanks. It was confusing me saying "it starts with a single cell". Really it starts with the whole line of white cells with the red cell in the middle, and the rules are applied to every continuous 3 cells at each iteration.
spencer_yie 3 days ago [-]
I am a bit fed up with this lad.
Out of the 6 comments 4 come straight from
The wolfram media web socket farm.
Come on guys. Please .
Rendered at 04:55:33 GMT+0000 (Coordinated Universal Time) with Vercel.
Just kidding, you know I'm here to talk about cellular automata.
Don't miss the "Personal Notes" section at the bottom. Neat history and insight into the process and life of a scientist.
> In a biological analogy we might have a genome that seems to lead to unbounded perhaps-tumor-like growth—but where actually the growth in the end “unexpectedly” stops.
I liked this aside. Early "cancer" detection can often lead to harm, mistaking benign for malicious growths.
> somehow follow from natural selection a general law...that would lead to the evolution of progressively more complex organisms, culminating in the “higher animals”. But absent the kind of model I’m discussing here, nothing in the later development of traditional evolutionary theory really successfully supported this—or was able to give much analysis of it.
A thing that comes to mind is Assembly Theory by Leroy Cronin and Sara Walker.
From there it's taking this and applying some pretty old genetic programming concepts to it to illustrate how we can see evolution in this every simple system. Wolfram has been very interested in cellular automata for a long time trying to find interesting ways to use it or study it even writing a gigantic book about it [1].
[0] https://content.wolfram.com/sites/43/2024/05/sw050124modelim...
[1] Called a new kind of science, he's made it available online if you're interested in learning about automata in maybe the deepest depth with some fun but ultimately not very useful theoretical applications: https://www.wolframscience.com/nks/
For example, the top left corner of the key shows that if a cell is blue and its neighbors are blue, it will remain blue. The one to the right shows that if a cell is blue and has a blue cell on the left and a red cell on the right, it will switch to white.
[1] https://www.wolframscience.com/nks/p24--how-do-simple-progra...