Self-Organising Textures
📰 Distill.pub
Researchers introduce Neural Cellular Automata (NCA) for generating self-organising textures, demonstrating its ability to learn diverse behaviors and solve complex tasks in a massively parallel and degenerate way
Action Steps
- Understand the concept of Neural Cellular Automata (NCA) and its application in generating self-organising textures
- Explore the inductive bias imposed by using cellular automata and its implications on task complexity
- Investigate the ability of NCA to learn diverse behaviors, such as generating stable images and segmenting images
- Analyze the degenerate nature of NCA and its impact on generalization to unseen situations
Who Needs to Know This
This research benefits AI engineers, ML researchers, and software engineers working on computer vision and pattern generation, as it provides a new approach to generating textures and understanding complex systems
Key Insight
💡 NCA's ability to impose a powerful inductive bias allows it to solve complex tasks in a degenerate and massively parallel way, making it a promising approach for computer vision and pattern generation
Share This
💡 Neural Cellular Automata (NCA) generates self-organising textures, learning diverse behaviors & solving complex tasks in a massively parallel way
Key Takeaways
Researchers introduce Neural Cellular Automata (NCA) for generating self-organising textures, demonstrating its ability to learn diverse behaviors and solve complex tasks in a massively parallel and degenerate way
Full Article
Published Time: Wed, 03 Aug 2022 01:57:28 GMT
# Self-Organising Textures
[Distill](https://distill.pub/)[About](https://distill.pub/about/)[Prize](https://distill.pub/prize/)[Submit](https://distill.pub/journal/)
# Self-Organising Textures
Neural Cellular Automata Model of Pattern Formation
Speed: 1x
  
Cell alignment
Rotation: 0 deg
 
Grid type
Textures
Inception
[interlaced_0172 (DTD)](https://www.robots.ox.ac.uk/~vgg/data/dtd/)
[Try in a Notebook](https://colab.research.google.com/github/google-research/self-organising-systems/blob/master/notebooks/texture_nca_tf2.ipynb)[Try in a Notebook](https://colab.research.google.com/github/google-research/self-organising-systems/blob/master/notebooks/texture_nca_pytorch.ipynb)
### Authors
### Affiliations
[Eyvind Niklasson](https://eyvind.me/)
[Google](https://ai.google/)
[Alexander Mordvintsev](https://znah.net/)
[Google](https://ai.google/)
Ettore Randazzo
[Google](https://ai.google/)
[Michael Levin](https://ase.tufts.edu/biology/labs/levin/)
[Tufts](https://tufts.edu/)
### Published
Feb. 11, 2021
### DOI
[10.23915/distill.00027.003](https://doi.org/10.23915/distill.00027.003)
### Contents
[Patterns, textures and physical processes](https://distill.pub/selforg/2021/textures#patterns-textures-and-physical-processes)
* [From Turing, to Cellular Automata, to Neural Networks](https://distill.pub/selforg/2021/textures#from-turing-to-cellular-automata-to-neural-networks)
* [NCA as pattern generators](https://distill.pub/selforg/2021/textures#nca-as-pattern-generators)
* [Related work](https://distill.pub/selforg/2021/textures#related-work)
[Feature Visualization](https://distill.pub/selforg/2021/textures#feature-visualization)
* [NCA with Inception](https://distill.pub/selforg/2021/textures#nca-with-inception)
[Other interesting findings](https://distill.pub/selforg/2021/textures#other-interesting-findings)
* [Robustness](https://distill.pub/selforg/2021/textures#robustness)
* [Hidden States](https://distill.pub/selforg/2021/textures#hidden-states)
[Conclusion](https://distill.pub/selforg/2021/textures#conclusion)
This article is part of the [Differentiable Self-organizing Systems Thread](https://distill.pub/2020/selforg/), an experimental format collecting invited short articles delving into differentiable self-organizing systems, interspersed with critical commentary from several experts in adjacent fields.
[Self-classifying MNIST Digits](https://distill.pub/2020/selforg/mnist/)[Adversarial Reprogramming of Neural Cellular Automata](https://distill.pub/selforg/2021/adversarial/)
Neural Cellular Automata (NCA We use NCA to refer to both _Neural Cellular Automata_ and _Neural Cellular Automaton_.) are capable of learning a diverse set of behaviours: from generating stable, regenerating, static images , to segmenting images , to learning to “self-classify” shapes . The inductive bias imposed by using cellular automata is powerful. A system of individual agents running the same learned local rule can solve surprisingly complex tasks. Moreover, individual agents, or cells, can learn to coordinate their behavior even when separated by large distances. By construction, they solve these tasks in a massively parallel and inherently degenerate Degenerate in this case refers to the [biological concept of degeneracy](https://en.wikipedia.org/wiki/Degeneracy_(biology)). way. Each cell must be able to take on the role of any other cell - as a result they tend to generalize well to unseen situations.
In this work, we appl
# Self-Organising Textures
[Distill](https://distill.pub/)[About](https://distill.pub/about/)[Prize](https://distill.pub/prize/)[Submit](https://distill.pub/journal/)
# Self-Organising Textures
Neural Cellular Automata Model of Pattern Formation
Speed: 1x
  
Cell alignment
Rotation: 0 deg
 
Grid type
Textures
Inception
[interlaced_0172 (DTD)](https://www.robots.ox.ac.uk/~vgg/data/dtd/)
[Try in a Notebook](https://colab.research.google.com/github/google-research/self-organising-systems/blob/master/notebooks/texture_nca_tf2.ipynb)[Try in a Notebook](https://colab.research.google.com/github/google-research/self-organising-systems/blob/master/notebooks/texture_nca_pytorch.ipynb)
### Authors
### Affiliations
[Eyvind Niklasson](https://eyvind.me/)
[Google](https://ai.google/)
[Alexander Mordvintsev](https://znah.net/)
[Google](https://ai.google/)
Ettore Randazzo
[Google](https://ai.google/)
[Michael Levin](https://ase.tufts.edu/biology/labs/levin/)
[Tufts](https://tufts.edu/)
### Published
Feb. 11, 2021
### DOI
[10.23915/distill.00027.003](https://doi.org/10.23915/distill.00027.003)
### Contents
[Patterns, textures and physical processes](https://distill.pub/selforg/2021/textures#patterns-textures-and-physical-processes)
* [From Turing, to Cellular Automata, to Neural Networks](https://distill.pub/selforg/2021/textures#from-turing-to-cellular-automata-to-neural-networks)
* [NCA as pattern generators](https://distill.pub/selforg/2021/textures#nca-as-pattern-generators)
* [Related work](https://distill.pub/selforg/2021/textures#related-work)
[Feature Visualization](https://distill.pub/selforg/2021/textures#feature-visualization)
* [NCA with Inception](https://distill.pub/selforg/2021/textures#nca-with-inception)
[Other interesting findings](https://distill.pub/selforg/2021/textures#other-interesting-findings)
* [Robustness](https://distill.pub/selforg/2021/textures#robustness)
* [Hidden States](https://distill.pub/selforg/2021/textures#hidden-states)
[Conclusion](https://distill.pub/selforg/2021/textures#conclusion)
This article is part of the [Differentiable Self-organizing Systems Thread](https://distill.pub/2020/selforg/), an experimental format collecting invited short articles delving into differentiable self-organizing systems, interspersed with critical commentary from several experts in adjacent fields.
[Self-classifying MNIST Digits](https://distill.pub/2020/selforg/mnist/)[Adversarial Reprogramming of Neural Cellular Automata](https://distill.pub/selforg/2021/adversarial/)
Neural Cellular Automata (NCA We use NCA to refer to both _Neural Cellular Automata_ and _Neural Cellular Automaton_.) are capable of learning a diverse set of behaviours: from generating stable, regenerating, static images , to segmenting images , to learning to “self-classify” shapes . The inductive bias imposed by using cellular automata is powerful. A system of individual agents running the same learned local rule can solve surprisingly complex tasks. Moreover, individual agents, or cells, can learn to coordinate their behavior even when separated by large distances. By construction, they solve these tasks in a massively parallel and inherently degenerate Degenerate in this case refers to the [biological concept of degeneracy](https://en.wikipedia.org/wiki/Degeneracy_(biology)). way. Each cell must be able to take on the role of any other cell - as a result they tend to generalize well to unseen situations.
In this work, we appl
DeepCamp AI