Learning Robust Dialog Policies in Noisy Environments

Understand

Modern virtual personal assistants provide a convenient interface for completing daily tasks via voice commands.

• An important consideration for these assistants is the ability to recover from automatic speech recognition (ASR) and natural language understanding (NLU) errors.
• In this paper, we focus on learning robust dialog policies to recover from these errors.
• To this end, we develop a user simulator which interacts with the assistant through voice commands in realistic scenarios with noisy audio, and use it to learn dialog policies through deep reinforcement learning.