Neural Cellular Automata
Exploring self-organizing systems and emergent pattern generation through biological-inspired computation.

This project delves into the fascinating world of Neural Cellular Automata (NCA), a class of models that learn to produce complex, dynamic patterns through local interactions. As part of my work with Monash DeepNeuron's advanced research division, I've been exploring how these systems, inspired by biological processes like morphogenesis, can model embryo development and generate intricate, life-like behaviors.

Core Features

Pattern Generation

Trained various NCA models to generate diverse, complex, and aesthetically pleasing patterns and textures that emerge from simple local rules and interactions.

Self-Organizing Systems

Explored emergent behavior where simple local cellular rules give rise to complex global order, mimicking biological morphogenesis and development processes.

Procedural Generation

Investigated applications for procedural content generation in games and simulations, creating dynamic environments and unique visual effects through cellular evolution.

Interactive Visualization

Built a real-time web-based visualizer using WebGL to observe the growth, evolution, and regenerative capabilities of cellular automata systems.

Technologies Used

PyTorchWebGLJavaScriptPythonGLSLC++