Deep Learning and Neural Networks

9.

Deep Reinforcement Learning

9.1.

Reinforcement Learning Fundamentals

9.1.1.

Agent-Environment Interaction

9.1.2.

States, Actions, and Rewards

9.1.2.1.

9.1.2.2.

9.1.2.3.

Reward Function

9.1.3.

Markov Decision Processes

9.1.3.1.

Markov Property

9.1.3.2.

Transition Probabilities

9.1.4.

Policies and Value Functions

9.1.4.1.

Policy Definition

9.1.4.2.

State Value Function

9.1.4.3.

Action Value Function

9.1.5.

Exploration vs. Exploitation

9.1.5.1.

Multi-Armed Bandit Problem

9.1.5.2.

Epsilon-Greedy Strategy

9.2.

Value-Based Methods

9.2.1.

Q-Learning Algorithm

9.2.1.1.

Temporal Difference Learning

9.2.1.2.

Q-Table Updates

9.2.2.

Deep Q-Networks (DQN)

9.2.2.1.

Neural Network Q-Function Approximation

9.2.2.2.

DQN Architecture

9.2.2.3.

Training Procedure

9.2.3.

DQN Improvements

9.2.3.1.

Experience Replay

9.2.3.1.1.

Replay Buffer Mechanism

9.2.3.1.2.

Breaking Temporal Correlations

9.2.3.2.

Target Networks

9.2.3.2.1.

Stable Target Values

9.2.3.2.2.

Periodic Updates

9.2.3.3.

9.2.3.4.

9.2.3.5.

Prioritized Experience Replay

9.3.

Policy-Based Methods

9.3.1.

Policy Gradient Theorem

9.3.2.

REINFORCE Algorithm

9.3.2.1.

Monte Carlo Policy Gradient

9.3.2.2.

Baseline Subtraction

9.3.3.

Actor-Critic Methods

9.3.3.1.

Actor and Critic Roles

9.3.3.2.

Advantage Function

9.3.3.3.

Asynchronous Advantage Actor-Critic (A3C)

9.3.4.

Proximal Policy Optimization (PPO)

9.3.4.1.

Clipped Surrogate Objective

9.3.4.2.

Trust Region Methods

9.4.

Model-Based Reinforcement Learning

9.4.1.

Environment Model Learning

9.4.2.

Planning with Learned Models

9.4.3.

Model-Predictive Control

9.5.

Multi-Agent Reinforcement Learning

9.5.1.

Cooperative Learning

9.5.2.

Competitive Learning

9.5.3.

Communication Protocols

9.6.

Applications of Deep Reinforcement Learning

9.6.1.

9.6.1.1.

9.6.1.2.

9.6.1.3.

Real-Time Strategy Games

9.6.2.

9.6.2.1.

Manipulation Tasks

9.6.2.2.

9.6.2.3.

9.6.3.

Autonomous Systems

9.6.3.1.

Autonomous Driving

9.6.3.2.

9.6.4.

Resource Management

9.6.5.

Trading and Finance

Previous

8. Generative Models

Go to top

Next

10. Advanced Topics and Specialized Architectures