I am currently trying to develop a reinforcement learning agent. I am particularly interested in applying the aforementioned project within the Rasa - Stack framework. That said, it’s not very clear to me if Rasa supports this kind of customization. More specifically, am I able to use Rasa NLU as it is and at the same time swap the interactive learning process, with my custom made reinforcement learning process, within the dialogue management of Rasa Core?
Cheers and thanks in advance