I believe algorithmic information theory will turn out to be key to physics, evolution, life, mind - and language, already. But don't yet see how to do it. > a

I believe algorithmic information theory will turn out to be key to physics, evolution, life, mind - and language, already. But don't yet see how to do it. > approximate the algorithmic information content of a network Haven't read the paper, but I'm not seeing how you could approximate it without throwing the baby out with the bathwater. > bias evolving programs toward simplicity [...] Rewarding something like shorter program lengths, for example Seems the simplest way to incorporate simplicity. However, e.g. some simple organisms have far longer DNA than us. And as for larger building blocks, e.g. cells, slime molds are unicellular with many features of higher life. The structure observed in nature is a byproduct of structures being a larger unit to navigate search space: like taking a highway, then main road, then side street to reach your destination. Or to home in on a decimal integer by varying any individual digit, instead of only adding +1 or -1 to the entire number. The search strategy modifies the probability density over the search space, rather than restricting it (or, explicity modifying the fitness function over the search space), yielding emergent structure.