NeuralKart - A Real-time Mario Kart AI using CNNs, Offline Search, and DAGGER

Set-up

The project has only been tested on Windows, but might work in other systems with adjustments.

The following Python dependencies need to be installed.

Our scripts are all written for the BizHawk emulator, which has embedded Lua scripting. To get BizHawk you first need to install the prerequisites - https://github.com/TASVideos/BizHawk-Prereqs/releases. Then you can download BizHawk and unzip it to any directory - https://github.com/TASVideos/BizHawk/releases/

TensorKart - The first MarioKart deep learning project, which we started from as our baseline.
Deep Learning for Real-Time Atari Game Play Using Offline Monte-Carlo Tree Search Planning - The idea for using a search-based AI for teaching the Convnet AI came from this paper.
A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning - The DAGGER algorithm was first introduced in this paper.
MarioKart 64 NEAT - This AI uses the NEAT algorithm to genetically evolve a shallow neural network
weatherton/BizHawkMarioKart64 - Some MarioKart 64 scripts which we used as a reference for memory locations.

Name		Name	Last commit message	Last commit date
Latest commit History 143 Commits
lualibs		lualibs
tools		tools
.gitignore		.gitignore
AxisTest.lua		AxisTest.lua
Demo.lua		Demo.lua
Evaluate.lua		Evaluate.lua
Play.lua		Play.lua
PlayAndSearch.lua		PlayAndSearch.lua
PlayAndSearch_new.lua		PlayAndSearch_new.lua
PlayAndSearchbeam.lua		PlayAndSearchbeam.lua
README.md		README.md
SearchAI.lua		SearchAI.lua
SearchAIbeam.lua		SearchAIbeam.lua
demo.gif		demo.gif
predict-server.py		predict-server.py
train.py		train.py
util.lua		util.lua