ConnectX RL Project

This repository contains a Reinforcement Learning (RL) project for the ConnectX game, leveraging advanced techniques to train an AI agent to play optimally.

Contents

• connectx_rl_project.ipynb: Jupyter Notebook with the full training and evaluation workflow.
• connectx_rl_project.html: HTML export of the notebook for quick preview.

Features

• Implements Dueling Deep Q-Network (DQN) architecture.
• Uses Kaggle's ConnectX environment for agent training and testing.
• Visualizes performance metrics and evaluation results.

How to Use

1. Clone this repository:

   git clone https://github.com/mnikoopayan/connectx-rl.git
   cd connectx-rl

2. Open the Jupyter Notebook (connectx_rl_project.ipynb) to view or run the code.
3. Submit the trained agent to Kaggle competitions using the provided files.

Project Highlights

-Efficient RL implementation with state-of-the-art techniques. -Detailed visualizations of training loss, win rates, and evaluation results. -Easy integration with Kaggle's ConnectX competition.