# μNCA: Texture Generation with Ultra-Compact Neural Cellular Automata

arXiv: https://arxiv.org/abs/2111.13545 · HTML: https://mlsanity.com/paper/2111.13545/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 40 references that are in the map.
- First posted on arXiv 26 Nov 2021.

## Builds on (most cited first, 15 of 15)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification](https://mlsanity.com/paper/1502.01852/) (2015)
- [Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distances](https://mlsanity.com/paper/1306.0895/) (2013)
- [Describing Textures in the Wild](https://mlsanity.com/paper/1311.3618/) (2013)
- [Reconciling modern machine learning practice and the bias-variance trade-off](https://mlsanity.com/paper/1812.11118/) (2018)
- [Computational Optimal Transport](https://mlsanity.com/paper/1803.00567/) (2018)
- [Deep Double Descent: Where Bigger Models and More Data Hurt](https://mlsanity.com/paper/1912.02292/) (2019)
- [Improved Texture Networks: Maximizing Quality and Diversity in Feed-forward Stylization and Texture Synthesis](https://mlsanity.com/paper/1701.02096/) (2017)
- [OT-Flow: Fast and Accurate Continuous Normalizing Flows via Optimal Transport](https://mlsanity.com/paper/2006.00104/) (2020)
- [Diversified Texture Synthesis with Feed-forward Networks](https://mlsanity.com/paper/1703.01664/) (2017)
- [A Survey on Optimal Transport for Machine Learning: Theory and Applications](https://mlsanity.com/paper/2106.01963/) (2021)
- [A Generative Model for Texture Synthesis based on Optimal Transport between Feature Distributions](https://mlsanity.com/paper/2007.03408/) (2020)
- [Texture Generation with Neural Cellular Automata](https://mlsanity.com/paper/2105.07299/) (2021)
- [Innovative Non-parametric Texture Synthesis via Patch Permutations](https://mlsanity.com/paper/1801.04619/) (2018)

## Built on by (most cited first, 13 of 13)

- [Attention-based Neural Cellular Automata](https://mlsanity.com/paper/2211.01233/) (2022)
- [DyNCA: Real-time Dynamic Texture Synthesis Using Neural Cellular Automata](https://mlsanity.com/paper/2211.11417/) (2022)
- [Mesh Neural Cellular Automata](https://mlsanity.com/paper/2311.02820/) (2023)
- [Arbitrarily Scalable Environment Generators via Neural Cellular Automata](https://mlsanity.com/paper/2310.18622/) (2023)
- [Neural cellular automata: applications to biology and beyond classical AI](https://mlsanity.com/paper/2509.11131/) (2025)
- [Learning spatio-temporal patterns with Neural Cellular Automata](https://mlsanity.com/paper/2310.14809/) (2023)
- [NoiseNCA: Noisy Seed Improves Spatio-Temporal Continuity of Neural Cellular Automata](https://mlsanity.com/paper/2404.06279/) (2024)
- [Infinite Texture: Text-guided High Resolution Diffusion Texture Synthesis](https://mlsanity.com/paper/2405.08210/) (2024)
- [AdaNCA: Neural Cellular Automata As Adaptors For More Robust Vision Transformer](https://mlsanity.com/paper/2406.08298/) (2024)
- [Multi-Texture Synthesis through Signal Responsive Neural Cellular Automata](https://mlsanity.com/paper/2407.05991/) (2024)
- [Future Research Avenues for Artificial Intelligence in Digital Gaming: An Exploratory Report](https://mlsanity.com/paper/2412.14085/) (2024)
- [Von Neumann Networks](https://mlsanity.com/paper/2605.05780/) (2026)
- [Applications of Neural Cellular Automata: State of the Art, Challenges and Opportunities](https://mlsanity.com/paper/2609.24595/) (2026)

## Abstract (arXiv)

We study the problem of example-based procedural texture synthesis using highly compact models. Given a sample image, we use differentiable programming to train a generative process, parameterised by a recurrent Neural Cellular Automata (NCA) rule. Contrary to the common belief that neural networks should be significantly over-parameterised, we demonstrate that our model architecture and training procedure allows for representing complex texture patterns using just a few hundred learned parameters, making their expressivity comparable to hand-engineered procedural texture generating programs. The smallest models from the proposed μNCA family scale down to 68 parameters. When using quantisation to one byte per parameter, proposed models can be shrunk to a size range between 588 and 68 bytes. Implementation of a texture generator that uses these parameters to produce images is possible with just a few lines of GLSL or C code.