Deep Reinforcement Learning for Multi-Domain Dialogue Systems

TL;DR

This paper introduces NDQN, a scalable deep reinforcement learning method for multi-domain dialogue systems, demonstrating improved performance over traditional DQN in simulation experiments for restaurant and hotel information-seeking tasks.

Contribution

The paper proposes NDQN, a novel multi-domain reinforcement learning approach that enhances scalability and effectiveness in dialogue policy learning.

Findings

NDQN outperforms DQN in scalability and dialogue management.

Experimental results show improved dialogue success rates.

NDQN is promising for real-world multi-domain dialogue systems.

Abstract

Standard deep reinforcement learning methods such as Deep Q-Networks (DQN) for multiple tasks (domains) face scalability problems. We propose a method for multi-domain dialogue policy learning---termed NDQN, and apply it to an information-seeking spoken dialogue system in the domains of restaurants and hotels. Experimental results comparing DQN (baseline) versus NDQN (proposed) using simulations report that our proposed method exhibits better scalability and is promising for optimising the behaviour of multi-domain dialogue systems.