I am trying to see if I understood the main points of this paper correctly. The background portion seemed delightfully obvious once it was pointed out to me:
- Unstable systems are advantageous from an evolutionary perspective because they allow rapid responses to environmental inputs.
- Unstable systems have to be carefully regulated because extreme responses frequently cause problems.
- Regulation is expensive, so it is a balancing act between maintaining responsive, unstable systems and maintaining sufficient regulation.
- Over time large systems are likely to fail because of the competing requirements of this balancing act.
Your work applies an information bottleneck iteration to two(?) previously established theorems and results in a model that at least approximates some interesting values or behaviors:
- You find that failure of regulation in unstable systems behaves mathematically like diffusion.
- You also find a proportionality for the minimal amount of regulation or maintenance needed to delay deterioration of an unstable system.
I understood the discussion that this could have broad implications across any subject where unstable control systems are important.
Did I correctly summarize this work?
And a second question: What kind of quantitative inputs would be required to test or apply the model? Are such inputs already available from experiment, or would they be very difficult to obtain?
As someone unfamiliar with the field, I don't expect detailed (or any) answers, but I found the topic intriguing. I will be looking up some of the referenced articles. Thank you.