Training an RL Agent to manage your emails
Email has become an integral part of many people's lives. The number of emails that an average working professional goes through in a workday is growing daily. While a lot of email filters exist for spam control, how nice would it be to have an intelligent Agent that can perform a series of email management tasks that just provide a task description (through text or speech via speech-to-text) and are not limited by any APIs that have rate limits? In this recipe, you will develop a deep RL Agent and train it on email management tasks! A set of sample tasks can be seen in the following image:
Figure 6.15 – A sample set of observations from the randomized MiniWoBEmailInboxImportantVisualEnv environment
Let's get into the details!
Getting ready
To complete this recipe, make sure you have the latest version. First, you will need to activate the tf2rl-cookbook Python/conda virtual environment. Make...
