1. Learn
    2. /
  2. Courses
    3. /
  3. Deep Reinforcement Learning in Python
  • 1

    Introduction to Deep Reinforcement Learning

    Free

    Discover how deep reinforcement learning improves upon traditional Reinforcement Learning while studying and implementing your first Deep Q Learning algorithm.

  • 2

    Deep Q-learning

    Dive into Deep Q-learning by implementing the original DQN algorithm, featuring Experience Replay, epsilon-greediness and fixed Q-targets. Beyond DQN, you will then explore two fascinating extensions that improve the performance and stability of Deep Q-learning: Double DQN and Prioritized Experience Replay.

  • 3

    Introduction to Policy Gradient Methods

    Learn about the foundational concepts of policy gradient methods found in DRL. You will begin with the policy gradient theorem, which forms the basis for these methods. Then, you will implement the REINFORCE algorithm, a powerful approach to learning policies. The chapter will then guide you through Actor-Critic methods, focusing on the Advantage Actor-Critic (A2C) algorithm, which combines the strengths of both policy gradient and value-based methods to enhance learning efficiency and stability.

  • 4

    Proximal Policy Optimization and DRL Tips

    Explore Proximal Policy Optimization (PPO) for robust DRL performance. Next, you will examine using an entropy bonus in PPO, which encourages exploration by preventing premature convergence to deterministic policies. You'll also learn about batch updates in policy gradient methods. Finally, you will learn about hyperparameter optimization with Optuna, a powerful tool for optimizing performance in your DRL models.

Initializing

Exercise

The Q-Network architecture

You are almost ready to train your first Deep Reinforcement Learning agent! Before you can go ahead with your first complete training loop, you need a neural network architecture to drive the agent's decisions and its ability to learn.

You will modify the generic architecture you defined in an earlier exercise. torch and torch.nn are imported into your exercises.

Instructions

100 XP
  • Instantiate the first hidden layer; its input will be the environment state, with dimension state_size.
  • Instantiate the output layer; it provides the Q-values for each action, with dimension action_size.
  • Complete the forward() method; use the torch.relu activation function for this example.