Learning Developmental Scaffoldings to Guide Self-Organisation

The paper introduces a novel model that integrates Neural Cellular Automata (NCA) with a learned coordinate-based pattern generator (SIREN) to explore the dynamics of self-organisation in biological systems. By jointly learning self-organisation rules and pre-patterns, the model demonstrates improved robustness, encoding capacity, and symmetry breaking compared to traditional self-organising methods. This research highlights the critical role of pre-patterns in shaping developmental processes, offering insights that could inform the design of more effective AI systems that leverage initial conditions for complex task completion.