Reinforcement Learning with TensorFlow

Summary

As a part of appendix, we covered a basic overview of continuous action space algorithms of the deep reinforcement learning topology, where we covered trust region policy optimization and deterministic policy gradients in brief. We also learned about the BLEU and ROUGE scores being actively used for evaluation in NLP-based sequential models. 

Finally, I would like to say that deep reinforcement learning is still a new topic as tons of more algorithms will be developed. But the most important thing that will help you to understand and explore those yet-to-be-discovered future algorithms will be the strong hold on the basics that this book has covered.