Deep Q-learning
Now that you have seen the basics of reinforcement learning and the most basic reinforcement learning model, Q-learning, it is time to move on to a more performant and more commonly used model called Deep Q-learning.
Deep Q-learning is a variant of Q-learning in which the Q-values are not just a list of expected Q-values for each combination of state and actions, updated by the Bellman equation. Rather, in Deep Q-learning, this estimation is done using a (deep) neural network.
If you are not familiar, neural networks are a class of machine learning models that are amongst the state of the art in terms of performance. Neural networks are largely used for many use cases in artificial intelligence, machine learning, and data science in general. Deep neural networks are the technology that allows many data science use cases such as Natural Language Processing (NLP), computer vision, and much more.
The idea behind the neural network is to pass an input data point...
