Greetings from Roll the Bones! Here we muse about complexity, learning, epistemology, accomplishing goals in complex environments, and whatever else I might be in the mood to discuss.
One thing I want to drive home is that life is not a game. The only simple, widely-understandable rules it follows are given by God Himself. In all other aspects, things are often more complex than they appear.
Personal Note
I’ve written a book! It’s an introductory text on coping with complexity and uncertainty. If you’re curious, please take a look. If you like it, please share. If you buy it, please read. If it helps you, please pay it forward and help someone else.
Gumroad: https://gum.co/notesoncomplexity
This Week’s Links
There are no affiliate links here, just things I’ve been reading. None of the authors have any idea their work is about to be featured.
Regression and Fire
In this three-blog-post series, Richard McElreath goes through “Causal Salad”, “Causal Design”, and “Full-Luxury Bayesian Inference” and how their naïve application (or near impossibility to achieve) leads to misunderstanding and error. These methods are widespread in the sciences. Perhaps(?), then, misunderstanding and error is widespread in the sciences. Perhaps it is lacking the proper mechanisms to check it.
“Researchers have been taught to think of statistical methods as a kind of sorcery that can conjure causal facts from data, as long as you have enough of it. Maybe this sounds harsh, but when a group of experts rely upon a set of methods to make decisions for them, but these experts have little mechanical understanding of the methods, fearing to deviate from convention, and lacking any formal framework for justifying these conventions, that sounds like sorcery. We must do better. “
My thoughts:
It is exciting to discover new things about the world - new laws and new causal relationships. Should it not also excite some of us to understand limits? To find boundaries? to understand the full shape and fragility of our assumptions about how things work?
It is frighteningly easy to use common statistical tools (including mutual information) to demonstrate correlations that we a priori know do not exist.
Inference tools (like AI) fail differently from humans. AI gets things hilariously wrong in ways a human never would, even though it can do some things that we struggle with. We see the world differently and think differently.
Finding Individuals with Information Theory
David Krakauer et al (U of Wisconsin, Max Planck Institute, Santa Fe Institute) decided they wanted an algorithmic, information-theory-centric way of identifying and distinguishing individuals from one another and their environment. To that end, they came up with their own answer to the question “What is an individual?”
My thoughts:
The environment is defined in here as something like “that which changes more slowly than individuals”. It acknowledges change, but otherwise is a bit vague. The environment, to my mind, is the set of processes that we don’t want to analyze as individuals at this time (they are not the target of analysis), but that’s just my knee-jerk reaction.
The authors acknowledge that individuality can be defined at different scales. For example, I am an individual, but I am made up of cells (and bacteria?) that may be considered individuals in their own right. At some level of resolution, my sub-components are made of things *not* considered individuals. Individuals can be “systems”, so long as they can be distinguished from their environment and other individuals by the information theoretic algorithms.
This is a remarkable step towards a more abstract, generalized way to define the things we want to study. When we define individuals manually, our hidden biases and assumptions may confound what we think we’re learning from our studies. Information theory can - perhaps - give us a chance to step back and re-assess what we think we know about the things we study.
Lenia - Continuous Cellular Automata
Lenia is a Cellular Automata (CA) system built by Bert Chan. It’s continuous, it’s pretty, and it’s got a lot of depth to it. CA have been a popular and fun way to play around with emergent and aesthetically pleasing behavior for a long time. This is a natural evolution, which will likely lead to a new wave of experiments and approaches.
My thoughts:
It’s pretty! Play with it in this web demo.
There is some depth here, and so there also seems to be a learning curve. I will need to invest some time to replicate notable patterns or discover my own - but that journey of discovery is rewarding in its own right. Are we not often delighted by beautiful things we didn’t know we could make?
Despite it’s lifelike appearance, this system does not operate on direct microscopic life analogues such as principles of survival, generational mutation/evolution, etc. For that, look to something like The Bibites or other similar simulations. Nevertheless, it certainly feels like there are secrets to be unlocked here.
Thank you for reading and engaging.
I appreciate you taking the time to read this newsletter. It’s free, but if you want to help support me you can always make a one time donation.
Engage with me on Twitter at @10101Lund about these or any other topics. If you find an error in this newsletter, please let me know and I’ll correct it. Archives of this newsletter are here.
Tell a friend. See you next issue!