Note

✅💯 it means, I have seen and tried them, I definitely recommend them to you

🔥 If you want to get free access to the books, courses and articles, get in touch with me with social media🔥

📣 You can help improve this repo, by giving me a ⭐star⭐ and ❤️following me❤️

Reinforcement learning (RL) is a field of machine learning concerned with how intelligent agents ought to take actions in an environment in order to maximize the notion of cumulative reward. Reinforcement learning is one of three basic machine learning paradigms, alongside supervised learning and unsupervised learning.

You can also read below blog to understand the key concepts in Reinforcement Learning

🔗 https://medium.com/intro-to-artificial-intelligence/key-concepts-in-reinforcement-learning-2af715dfbfa

Warning This is only a segguestion roadmap

*"Reinforcement Learning: An Introduction - Second Edition - Richard S. Sutton and Andrew G. Barto"

"Deep Reinforcement Learning with Python - Second Edition - Sudharsan Ravichandiran"

"Grokking Deep Reinforcement Learning - Miguel Morales"

"PyTorch 1.x Reinforcement Learning Cookbook - Yuxi (Hayden) Liu"

"Deep Reinforcement Learning Hands-On - 2nd Edition - Maxim Lapan"

"TensorFlow 2 Reinforcement Learning Cookbook - 2nd Edition - Praveen Palanisamy"

1958 Oct: Rosenblatt, F. "The perceptron: A probabilistic model for information storage and organization in the brain.."

1983 Oct: Andrew G. Barto, et al. "Neuronlike adaptive elements that can solve difficult learning control problems."

1988 Feb (TD): Sutton, R.S. "Learning to Predict by the Methods of Temporal Differences."

1992 May: Watkins, C.J.C.H., Dayan, P. "Q-learning". Mach Learn 8, 279–292 (1992) 

1994 Nov: G. A. Rummery, M. Niranjan. "On-Line Q-Learning Using Connectionist Systems"

1995 Mar: Gerald Tesauro. "Temporal Difference Learning and TD-Gammon."

2005 Oct: Riedmiller, Martin. "Neural fitted Q iteration–first experiences with a data efficient neural reinforcement learning method." 

2012 Jul (ALE): Bellemare, Marc G., et al. "The Arcade Learning Environment: An Evaluation Platform for General Agents." 

2013 Dec (DQN): Volodymyr Mnih, et al. "Playing Atari with Deep Reinforcement Learning." 

2015 Feb (DQN): Mnih, Volodymyr, et al. "Human-level control through deep reinforcement learning."

2015 Feb: Bernhard Schölkopf. "Learning to see and act." Nature518, pages486–487 (2015)

2015 Sep (DDQN): Hado van Hasselt, et al. "Deep Reinforcement Learning with Double Q-learning." 

2015 Sep (DDPG): Lillicrap, Timothy P., et al. "Continuous control with deep reinforcement learning." 

2015 Nov: Ziyu Wang, et al. "Dueling Network Architectures for Deep Reinforcement Learning." 

2015 Nov (PER): Schaul, Tom, et al. "Prioritized Experience Replay." 

2016 Jan: David Silver, et al. "Mastering the game of Go with deep neural networks and tree search."

2016 Jun: Brockman, Greg, et al. "Openai gym." arXiv preprint arXiv:1606.01540. 

2016 Feb (A3C): Volodymyr Mnih, et al. "Asynchronous Methods for Deep Reinforcement Learning." 

2017 Aug (PPO): Schulman, John, et al. "Proximal policy optimization algorithms." 

2017 Oct: Hessel, Matteo, et al. "Rainbow: Combining Improvements in Deep Reinforcement Learning." 

2018 Dec (SAC): Haarnoja, Tuomas, et al. "Soft Actor-Critic Algorithms and Applications." 

2020 Oct: Xing Wua, et al. "Adaptive stock trading strategies with deep reinforcement learning methods."

2022 Mar: Wen, Y. et al. "Improving the exploration efficiency of DQNs via the confidence bound methods."

1. Reinforcement Learning Specialization

Master the Concepts of Reinforcement Learning. Implement a complete RL solution and understand how to apply AI tools to solve real-world problems.

• Offered By UNIVERSITY OF ALBERTA ALBERTA MACHINE INTELLIGENCE INSTITUTE
• There are 4 Courses in this Specialization
  🔗Fundamentals of Reinforcement Learning
  🔗Sample-based Learning Methods
  🔗Prediction and Control with Function Approximation
  🔗A Complete Reinforcement Learning System (Capstone)

2. Become a Deep Reinforcement Learning Expert

Learn the deep reinforcement learning skills that are powering amazing advances in AI. Then start applying these to applications like video games and robotics.

• ✅💯Presented by UDACITY - Nanodegree Program

3. DeepMind x UCL | Deep Learning Lecture Series 2021

The Deep Learning Lecture Series is a collaboration between DeepMind and the UCL Centre for Artificial Intelligence.

• ✅💯Presented by DeepMind & UCL

4. Advanced AI: Deep Reinforcement Learning in Python

Created by Lazy Programmer Team, Lazy Programmer Inc.

• The Complete Guide to Mastering Artificial Intelligence using Deep Learning and Neural Networks

5. Deep Reinforcement Learning, Decision Making and control

Lectures for UC Berkeley CS 285: Deep Reinforcement Learning for Fall 2021 - Instructor: Sergey Levine

🔗 An introduction to Deep Q-Learning: let’s play Doom
🔗 Deep Reinforcement Learning: Pong from Pixels ✅💯Andrej Karpathy blog

🔗 Q-learning: a value-based reinforcement learning algorithm
🔗 Deep Q Network(DQN)- Applying Neural Network as a functional approximation in Q-learning
🔗 Diving deeper into Reinforcement Learning with Q-Learning

🔗 REINFORCE — a policy-gradient based reinforcement Learning algorithm
🔗 Policy Based Reinforcement Learning, the Easy Way
🔗 Reinforcement Learning Policy for Developers
🔗 An introduction to Policy Gradients with Cartpole and Doom

🔗 Introduction to Actor Critic in Reinforcement Learning
🔗 The Actor-Critic Reinforcement Learning algorithm

🔗 Getting Started With OpenAI Gym: The Basic Building Blocks

🔗 This Factory Robot Learns a New Job Overnight
🔗 RL — Basics algorithms and terms
🔗 The Complete Reinforcement Learning Dictionary

Note In this environment, the observation is an RGB image of the screen, which is an array of shape (210, 160, 3) Each action is repeatedly performed for a duration of k frames, where k is uniformly sampled from {2, 3, 4}. Our Target is to maximize our score.

Note Frozen lake involves crossing a frozen lake from Start(S) to Goal(G) without falling into any Holes(H) by walking over the Frozen(F) lake. The agent may not always move in the intended direction due to the slippery nature of the frozen lake.

Note

Note

Note

Note This recipe will walk you through the steps to build a Gridworld learning environment to train RL agents. Gridworld is a simple environment where the world is represented as a grid. Each location on the grid can be referred to as a cell. The goal of an agent in this environment is to find its way to the goal.