What is Blob Darwinism

Blob Darwinism is an in progress machine learning project that simulates an ecosystem of small beings called blobs. Blobs seek to maximize their survival by moving and eating food or other blobs. The behavior of each blob is governed by a unique neural network which is initially random. After a generation of blobs is simulated, a subsequent generation is created by selectively inheriting genetics from the best performers of the previous generation. The simulation is rendered using WebGL to allow for smooth fast animations.

Neural Network

This project uses a neural network written in vanilla JavaScript. This allows the blobs naturally evolve optimal behavior

Settings

The user can change how the simulation is run for different results. The settings that can currently be changed are:

  • When To End a Generation
  • Number of Starting Blobs
  • Number of Starting Food
  • Neural Network Hidden Layers
  • Neural Network Neurons
  • Mutation Rate
  • Mutation Range

Email

garrettjtongue@gmail.com

Phone

(917) 224-2939