Programmable Cellular Automata: readable code rules evolved via genetic programming
Amidos2006 · x · 2026-09-09
The author introduces Programmable Cellular Automata (PCA): classic CA rules are painful lookup tables and neural CAs are black boxes — PCA makes every rule readable code.
- CA is modularized into small programs: local functions (neighborhood → value), optional global functions (whole board → value, breaking strict locality so tasks like counting converge fast), and a decision function that combines outputs into the next cell state.
- Since rules are programs, they can be evolved with genetic programming: the chromosome is the list of all local/global/decision functions, crossover swaps programs between parents, mutation replaces a program.
- PCA was tested as a level generator on 3 PCG Benchmark problems (Binary maze, Zelda, Sokoban) with a cascaded fitness: levels must be 100% playable before optimizing diversity.
Related event: Programmable Cellular Automata: readable code replaces lookup tables(2 posts)→
More from Research
- If human insight remains an ineliminable complement, it will limit AI's gains in science — erikphoel · 2026-09-10
- Simplex launches research blog Belief Updates, showing transformer belief geometry forms telescoping cones — rishabh16_ · 2026-09-10
- After Navier-Stokes: the anti-hype take on AI's role in the math breakthrough — Jsevillamol · 2026-09-10
- DEER-3D fixes 3D-LLM grounding biases with error-driven counterfactual scene editing — YonatanBitton · 2026-09-10
- Universities' official policies contain straight-up misinformation about AI detectors — paulnovosad · 2026-09-10
- LangChain's Chase: Agent Gains Come from Harness, Not Model — Paper Shows 14pt Lift — hwchase17 · 2026-09-10