Domain-independent User Simulation with Transformers for Task-oriented Dialogue Systems
This work addresses the need for scalable and adaptable user simulators in dialogue policy optimization, reducing the time and cost of training with real users, though it is incremental as it builds on existing transformer methods.
The authors tackled the problem of domain-dependent user simulators for task-oriented dialogue systems by proposing a domain-independent transformer-based user simulator (TUS), which achieved competitive performance with rule-based simulators on pre-defined domains and generalized to unseen domains in a zero-shot fashion.
Dialogue policy optimisation via reinforcement learning requires a large number of training interactions, which makes learning with real users time consuming and expensive. Many set-ups therefore rely on a user simulator instead of humans. These user simulators have their own problems. While hand-coded, rule-based user simulators have been shown to be sufficient in small, simple domains, for complex domains the number of rules quickly becomes intractable. State-of-the-art data-driven user simulators, on the other hand, are still domain-dependent. This means that adaptation to each new domain requires redesigning and retraining. In this work, we propose a domain-independent transformer-based user simulator (TUS). The structure of our TUS is not tied to a specific domain, enabling domain generalisation and learning of cross-domain user behaviour from data. We compare TUS with the state of the art using automatic as well as human evaluations. TUS can compete with rule-based user simulators on pre-defined domains and is able to generalise to unseen domains in a zero-shot fashion.