Genetic Algorithms in Go
summary
Implements genetic algorithms in Go with concrete code on GitHub: DNA is byte arrays (text) or RGBA images, fitness is a normalized distance (RGBA per-pixel using repeated Pythagorean sums), selection uses a breeding pool with copy-weighted entries, crossover at a random midpoint and per-generation mutation. It evolves the phrase to be or not to be in under a few seconds (sometimes ~0.5s) versus a 10^? impossibility by brute force, and progressively approximates a 100×64 Mona Lisa image (fitness dropping from ~16k toward ~7k).
AI-written summary · about
chapters (9)
- 0:00Introduction
- 8:30Defining the organism
- 17:00Finding the organism
- 17:52Finding the difference
- 18:38Using the pigs trick
- 19:40Breeding pool mechanism
- 20:33Generic Algorithm
- 21:51Fitness
- 23:17Questions
related talks
24mGenetic Algorithms in Go
24m"Discovery of Worst Case Go Sort Inputs"
04mGo, C++ or Java for DNA sequencing?
51mBuilding Generative AI Applications in Go- Gari Singh, Google (Part 2)
22mCreating Immersive Generative Art With Go
37mA Funny Thing Happened On The Way To Reimplementing AlphaGo
18mMachine Learning and Go
01h 01mBuilding Generative AI Applications in Go- Gari Singh, Google (Part 1)