Reinforcement Learning of Multi-Domain Dialog Policies Via Action Embeddings
This reduces data needs for real-world dialog systems, but it is incremental as it builds on existing multi-domain learning approaches.
The paper tackles the problem of high data requirements for learning task-oriented dialog policies via reinforcement learning by leveraging cross-domain data and learning domain-agnostic action embeddings, resulting in a 35% reduction in dialogs needed and higher proficiency compared to separate domain policies.
Learning task-oriented dialog policies via reinforcement learning typically requires large amounts of interaction with users, which in practice renders such methods unusable for real-world applications. In order to reduce the data requirements, we propose to leverage data from across different dialog domains, thereby reducing the amount of data required from each given domain. In particular, we propose to learn domain-agnostic action embeddings, which capture general-purpose structure that informs the system how to act given the current dialog context, and are then specialized to a specific domain. We show how this approach is capable of learning with significantly less interaction with users, with a reduction of 35% in the number of dialogs required to learn, and to a higher level of proficiency than training separate policies for each domain on a set of simulated domains.