Nate GaylinnI love this paper! It's just the sort of experiment I'd like to run myself, and there's a chance I'll get to collaborate with Risi's lab on something like this.
I think this sort of research helps advance our understanding of evolution and intelligence. The key point here is that DNA is not "the program for" an organism. Rather, it's just one part in a system of constraints that shape the behavior of individual cells, which coordinate to make a body.
What isn't well appreciated is how that change in perspective affects evolution and the production of intelligent behavior! This paper explores the initial conditions that precede development. They show how patterning constraints imposed by a mother can make development more robust, and guide a growing embryo toward one of several possible adult forms. They speculate (and I agree) that such a system ought to make organisms more evolvable, too, though they haven't shown that yet.
Learning Developmental Scaffoldings to Guide Self-Organisation
From subcellular structures to entire organisms, many natural systems generate complex organisation through self-organisation: local interactions that collectively give rise to global structure without any blueprint of the outcome. Yet a significant portion of the information driving such processes is not produced by self-organisation itself, instead, it is often offloaded to initial conditions of the system. Biological development is a prime example, where maternal pre-patterns encode positional and symmetry-breaking information that scaffolds the self-organising process. From maternal morphogen gradients in early embryogenesis to tissue-level morphogenetic pre-patterns guiding organ formation, this transfer of information to initial conditions, analogous to a memory-compute trade-off in computational systems, is a fundamental part of developmental processes. In this work, we study this offloading phenomenon by introducing a model that jointly learns both the self-organisation rules and the pre-patterns, allowing their interplay to be varied and measured under controlled conditions: a Neural Cellular Automaton (NCA) paired with a learned coordinate-based pattern generator (SIREN), both trained simultaneously to generate a set of patterns. We provide information-theoretic analyses of how information is distributed between pre-patterns and the self-organising process, and show that jointly learning both components yields improvements in robustness, encoding capacity, and symmetry breaking over purely self-organising alternatives. Our analysis further suggests that effective pre-patterns do not simply approximate their targets; rather, they bias the developmental dynamics in ways that facilitate convergence, pointing to a non-trivial relationship between the structure of initial conditions and the dynamics of self-organisation.