Flappy Bird WASM Neuro Evolution

Thanks for visiting!

This is an implementation of the famous game Flappy Bird. It is written in Rust and compiled to WebAssembly. Click here to see the source code.

You can visualize AIs trying to learn the game. The algorithm is based on the papers Evolving Neural Networks through Augmenting Topologies and Gated Recurrent Units (GRU).

You can also play by selecting Give me the black bird for next generation. You can always decide to opt in or out.

The goal of this demo is to showcase my library neat-gru-rust

There are two version of the game, classic and aerodynamics. The difference is that in classic mode, the fall velocity is reset at every jump. Whereas in aerodynamics, jumping only increments the current falling speed; making it much harder to control.

The inputs passed are:

Game Parameters