Q-Learning based Reinforcement Learning implementation, make AI self-learn to play Cartpole and 3 Atari games (Boxing, Pong, Pacman)

An implementation of the Atari game Breakout, where an AI plays the game using Deep Q-Network (DQN) learning.

An AI designed to run Atari games using Q-Learning. Based on paper 'Playing Atari with Deep Reinforcement Learning' by Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Alex Graves, Ioannis Antonoglou, Daan Wierstra, and Martin Riedmiller.

🕹️ Train an AI actor to play the classic Atari game 'Breakout' using Deep Q Learning.

Train an AI agent to play Atari Breakout using Dueling Deep Q-Networks (Dueling DQN) in PyTorch. This project implements deep reinforcement learning techniques like experience replay, target networks, and frame stacking to improve gameplay performance.

BreakoutAI is an exciting project dedicated to conquering the classic Atari Breakout game through the power of reinforcement learning. Leveraging the state-of-the-art Stable Baselines3 library, our AI agent, armed with a Deep Q-Network (DQN), undergoes intense training sessions to master the art of demolishing bricks.

Q-learning script that can learn to play any Atari game using open AI gym.

Reinforcement Learning ⬩ Computer Vision ⬩ Taught an AI how to play a basic shooting match using Deep Q-Learning. Inspired by @simoninithomas's DQN implementation in Atari Space Invaders 2600.

We introduce the first deep reinforcement learning agent that learns to beat Atari games with the aid of natural language instructions. The agent uses a multimodal embedding between environment observations and natural language to self-monitor progress through a list of English instructions, granting itself additional reward for completing instructions in addition to increasing the game score. Our agent significantly outperforms Deep-Q Networks, Asynchronous Advantage Actor-Critic (A3C) agents, and the best agents posted to Open-AI Gym on what is often considered the hardest Atari 2600 environment: MONTEZUMA’S REVENGE.

David Silver FRS (born 1976) leads the reinforcement learning research group at DeepMind and was lead researcher on AlphaGo, AlphaZero and co-lead on AlphaStar. He graduated from Cambridge University in 1997 with the Addison-Wesley award, and befriended Demis Hassabis whilst there.[1] Subsequently, Silver co-founded the video games company Elixir Studios, where he was CTO and lead programmer, receiving several awards for technology and innovation.[1][2] Silver returned to academia in 2004 at the University of Alberta to study for a PhD on reinforcement learning, where he co-introduced the algorithms used in the first master-level 9×9 Go programs.[3][4] His version of program MoGo (co-authored with Sylvain Gelly) was one of the strongest Go programs as of 2009.[5] Silver was awarded a Royal Society University Research Fellowship in 2011, and subsequently became a lecturer at University College London, where he is now a professor.[6] His lectures on Reinforcement Learning are available on YouTube.[7] Silver consulted for DeepMind from its inception, joining full-time in 2013. His recent work has focused on combining reinforcement learning with deep learning, including a program that learns to play Atari games directly from pixels.[8] Silver led the AlphaGo project, culminating in the first program to defeat a top professional player in the full-size game of Go.[9] AlphaGo subsequently received an honorary 9 Dan Professional Certification; and won the Cannes Lion award for innovation.[10] He then led development of AlphaZero, which used the same AI to learn to play Go from scratch (learning only by playing itself and not from human games) before learning to play chess and shogi in the same way, to higher levels than any other computer program. Silver is among the most published members of staff at DeepMind, with over 67,000 citations and has an h-index of 66.[11] He was awarded the 2019 ACM Prize in Computing for breakthrough advances in computer game-playing.[12] In 2021, Silver was elected Fellow of the Royal Society for his contributions to Deep Q-Networks and AlphaGo.

AI that plays Atari Pong using deep q learning

No description available

No description available

Deep Q-Learning AI's to play Atari OpenAI Gym Games

Creating a Deep Q Learning Reinforcement Learner for playing Atari games (using AI Gym)

Implementation of Deep Q Learning on open AI gym cartpole and atari breakout environments

An implementation of the Atari game Breakout, where an AI plays the game using Deep Q-Network (DQN) learning.

Training a central Q-learning (super brain agent) for a cooperative multi-agent reinforcement learning (MARL) game from the PettingZoo library Atari environments to develop AI agents for Atari 2600 games. T

An AI agent trained to play Space Invaders using Reinforcement Learning. This project demonstrates how Deep Q-Learning (DQN) and other RL techniques can be applied to master classic Atari games.

Projet d’IA où j’ai entraîné des agents sur plusieurs jeux Atari en combinant Q-Learning et Sarsa, explorant l’apprentissage par renforcement pour optimiser la prise de décision et améliorer les performances de jeu

This research aims to explore the field of Multi-Task Reinforcement Learning. With a slight modification to the Deep Q-Network algorithm, Polygame AI is capable of learning to play several Atari games. Future work involves scaling our algorithm up to a more complex and diverse set of tasks.