user-guide on Deep-RL

[TimotheeMathieu]

Proposition of outline for Deep rl part of user-guide:

• Deep RL experimental design with rlberry
• Accessing the neural networks of an agent
• hyperparameter optimization in rlberry

EDIT: this part can also be a place to add a template for a typical workflow for Deep RL research following for instance the workflow of https://arxiv.org/abs/2304.01315