deep-q-networks

This repository hosts the implementation of autonomous vehicle navigation using RL techniques, with a specific emphasis on Deep Q-Networks (DQN) and Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithms. We focus on training a TurtleBot3 robot to navigate autonomously through environments while intelligently avoiding moving obstacles.

A TensorFlow attempt to reimplement the IEEE VTC2023-Spring paper “Deep Reinforcement Learning-Based Resource Allocation for Cellular V2X Communications”

RL4CAD: Personalized Decision Making for Coronary Artery Disease Treatment using Offline Reinforcement Learning

This collection features a wide array of simple implementations of machine learning algorithms spanning various methodologies and applications.

This project aims to experiment with a car game, exploring whether a DQN (Deep Q Network) agent can autonomously learn to play the game.

Deep Reinforcement Learning for solving discrete environments

Implementation wumpus with Deep-Q Networks

Deep Q-Network (DQN) agent playing ViZDoom from raw pixels. Features custom composite actions, shaped rewards, MLflow logs, and a retro HUD dashboard served on Oracle Cloud VM.

🐦A DQN agent that learned to play Flappy Bird from scratch . Built with PyTorch, experience replay, and target networks. Trained over 100K+ episodes on GPU. Includes pre-trained model, training logs, and the full optimization story of going from zero convergence to 101.9 reward.

This project is a practical exercise to implement and test a reinforcement learning agent using the CartPole environment from OpenAI Gym. The goal is to develop an agent capable of maintaining a pole balanced on a moving cart by applying deep reinforcement learning techniques.