Neural Cellular Automata Learn General Features in their Hidden Channels

Neural Cellular Automata (NCAs) outperform comparable recurrent and feed-forward architectures on few-shot and scale-variant MNIST benchmarks with a parameter budget of roughly 9,800 parameters, according to a paper published on arXiv as 2609.21870v1. The paper's mechanistic analysis found that NCA hidden channels decouple feature extraction from uniform classification consensus by absorbing morphological complexity and converging to mutually orthogonal states, capturing general scale-invariant topological primitives rather than class-specific templates. A student model achieved strong few-shot performance on unseen classes using features transferred from a teacher trained only on digits 0-5, which the authors present as a decentralized computational substrate for parameter-efficient transfer learning.

arXiv:2609.21870v1 Announce Type: new Abstract: Modern deep learning models achieve impressive generalization through over-parameterization, but this paradigm often struggles with overfitting and memorization in few-shot regimes. Neural Cellular Automata NCAs offer a highly parameter-efficient alternative, yet research has focused primarily on their output, leaving the role of their internal hidden channels largely unexplored. In this paper, we investigate the internal dynamics of NCA hidden channels and introduce a novel transfer-learning mechanism that injects a pretrained teacher's hidden states into a student model to guide early optimization. Evaluated on few-shot and scale-variant MNIST benchmarks, NCAs outperform comparable recurrent and feed-forward architectures, demonstrating superior generalization with a minimal parameter budget ~9,800 parameters . Mechanistic analysis reveals that the hidden channels decouple feature extraction from uniform classification consensus by absorbing morphological complexity and converging to mutually orthogonal states. Furthermore, we demonstrate that these hidden channels capture general, scale-invariant topological primitives rather than class-specific templates. This allows a student model to achieve strong few-shot performance on unseen classes using features transferred from a teacher trained only on a subset of digits 0-5 . Our results highlight the potential of utilizing hidden-state dynamics as a robust, decentralized computational substrate for parameter-efficient transfer learning