Neural Cellular Automata
Overview
Neural cellular automata study how many simple cells, each carrying a state and following a shared learned update rule, can produce complex behavior through repeated local interaction. Researchers use them for self-organizing growth and repair, image and texture generation, spatial reasoning, and simulations of physical dynamics such as fluids, fires, and traffic. Common approaches combine differentiable neighborhood operations with recurrent updates, sometimes adding stochasticity, physical constraints, attention, or meta-learning to improve robustness and transfer. A central challenge is balancing expressive behavior and generalization with long-term stability, since a system that matches a target briefly may fail under perturbations or extended rollouts.
Latest papers
39 papers3 papers in the last four weeks, down 62% on the four weeks before. 0.0% of all new papers.
Distributed Swarm Morphogenesis with Neural Cellular Automata
Multiclass Boundary Refinement with Neural Cellular Automata for Remote Sensing Land Cover Segmentation
Learning Traffic Flow Dynamics with Stochastic Physics-Informed Neural Cellular Automata
2D Spatial Reasoning with Adaptive Neural Cellular Automata
MetaLearnNCA: Few-Shot Offline Meta-Learning via Interacting Neural Cellular Automata
Where Does Randomness Matter in Neural Cellular Automata?
Reasoning with Neural Cellular Automata
On Growth and Form, and Function: Reusable Regulatory Handles Control Phenotypic Variation
Online Task Adaptation via Self-Organisation
Applications of Neural Cellular Automata: State of the Art, Challenges and Opportunities
Neural Cellular Automata Learn General Features in their Hidden Channels
Self-Replicating Neural Cellular Automata: Quantifying Emergent Phenotypic and Genotypic Diversity in an OpenEnded Substrate
Learning PDE Time-Stepping with Neural Cellular Automata
Coarse to Fine: Iterative Adversarial Neural Cellular Automata for Medical Image Synthesis
Sterilizable Scene Graph Generation for Operating Rooms
TextNCA: Neural Cellular Automata for Language Modeling via Hierarchical Local Attention
Emulating Cosmic Structure Formation with a Lagrangian Neural Cellular Automaton
Transient State Reorganization and Cell Differentiation in the Developmental Dynamics of Growing Neural Cellular Automata
One-Shot Generative Design for Disordered Metamaterials via Self-Organizing Neural Cellular Automata
Structured Fluctuations and the Information Dynamics of Self-Maintenance in Growing Neural Cellular Automata
Architecture Generalization with MetaNCA
Mass Conservation as an Inductive Bias for Self-Organized Criticality in NCA Reservoirs
Communication Heterogeneity and Collective Consensus in Neural Cellular Automata
languages'' as sub-populations that read one another's messages through a translation with a tunable linguistic distance''. We find that linguistic distance slows consensus, that it produces mild divergence between groups rather than full fragmentation, and that a collective whose shared rule was trained under diverse protocols is robust to mismatch; a homogeneously trained one is not. The findings hold on both a ring and a two-dimensional grid, and admit a natural reading as Ising relaxation, in which a foreign-language region acts as a boundary defect that leaves the system in a higher-energy, partially ordered state. These patterns are qualitatively consistent with effects reported in human group studies, suggesting that distance between communication protocols is a minimal mechanism sufficient to produce them, without anything language-specific.