Skip to content

Train Your Foes Logo

Welcome to the Train Your Foes Documentation

This site contains the complete documentation for Train Your Foes, a game project that merges classic 2D platforming with an adaptive AI powered by Reinforcement Learning. Here you will find everything from how to play the game to a deep dive into the technology and development process behind it.

What is Train Your Foes?

Train Your Foes is a proof-of-concept game built in the Unity Engine. It is designed as a two-part experience:

  1. The Gauntlet: Four levels of intense, precision platforming designed to test player skill and mastery of the game's mechanics.
  2. The Duel: A final, turn-based boss battle against an AI that does not follow a predictable script. Its strategies are the result of a custom-built Q-Learning model, creating a unique and challenging encounter.

This project is a practical exploration of applying machine learning concepts to create more dynamic and engaging opponents in video games.

This documentation is structured to guide you, whether you're a player, a developer, or an AI enthusiast. Use the navigation bar at the top, or start with one of the key sections below.

  • Getting Started & Gameplay: Learn how to install, run, and play the game.
  • Development Deep Dive: Explore the complete story of the project, from the initial learning and challenges to the final implementation of the game and its AI.

Get Started

Choose a path below to begin your exploration.

For Players »

New to the project and just want to play? This is your starting point. Learn how to download and run the game on your system.

View the Getting Started Guide

For Developers & The Curious »

Interested in the story behind the game? This section details the entire development lifecycle, including the initial learning, the challenges faced with ML-Agents, and the pivot to a custom AI solution.

Read the Development Journey

For AI Enthusiasts »

Want to see the technical details? This section provides a direct look at the AI's architecture, including its State Space, Action Space, and the Reward Function that drives its learning.

Analyze the AI Model