HyperNCA: Growing Developmental Networks with Neural Cellular Automata
This work addresses the challenge of developing more biologically plausible and adaptable neural networks for reinforcement learning, though it appears incremental in applying neural cellular automata to network growth.
The authors tackled the problem of growing artificial neural networks through a self-organized developmental process, inspired by biological systems, and demonstrated that their HyperNCA method can grow networks capable of solving common reinforcement learning tasks, including transformations for task variations.
In contrast to deep reinforcement learning agents, biological neural networks are grown through a self-organized developmental process. Here we propose a new hypernetwork approach to grow artificial neural networks based on neural cellular automata (NCA). Inspired by self-organising systems and information-theoretic approaches to developmental biology, we show that our HyperNCA method can grow neural networks capable of solving common reinforcement learning tasks. Finally, we explore how the same approach can be used to build developmental metamorphosis networks capable of transforming their weights to solve variations of the initial RL task.