Investigating Reinforcement Learning for Communication Strategies in a Task-Initiative Setting

Baber Khalid, Matthew Stone

Many conversational domains require the system to present nuanced information to users. Such systems must follow up what they say to address clarification questions and repair misu…