Deep reinforcement learning

In order for the reinforcement learning algorithm to be deployed in real-world use cases and scenarios, we need to leverage the power of deep neural networks, which can infer the information from the environments in a human-like manner. One of the goals of AI is to augment human capabilities by creating autonomous agents that interact with the environment in which they operate, learn optimal behaviors that improve over time, and learn from mistakes.

For example, the signals from the video camera can be interpreted using a deep neural network. Once this signal is interpreted, the objects and patterns observed by the camera can be analyzed with the help of a deep neural network, as we have seen in the chapters on artificial neural networks (ANNs). These deep neural networks can then be used for application of reinforcement learning algorithms for creating a navigation system that learns over a period of time based on the training feeds.

Fundamentally, a combination...