Neural Cellular Automata Achieve Arbitrary Resolution Through Implicit Decoding

This work solves a critical limitation of Neural Cellular Automata by pairing coarse-grid computation with lightweight implicit decoding, enabling real-time high-resolution synthesis while preserving the biological self-organizing properties that make NCAs uniquely robust and regenerative. By decoupling the computational lattice from the output resolution, the approach dramatically reduces memory and compute requirements while maintaining full parallelizability. This breakthrough opens new frontiers for generative design, texture synthesis, and morphogenetic modeling at previously unattainable scales.