Adaptive Natural Language Generation for Task-oriented Dialogue via Reinforcement Learning
This addresses the need for more robust and user-adaptive natural language generation in real-world dialogue systems, though it is incremental as it builds on existing RL methods for language generation.
The paper tackles the problem of generating adaptive utterances in task-oriented dialogue systems to handle environmental noise and varying user understanding levels, proposing ANTOR which uses reinforcement learning with an NLU module for rewards, and shows it can generate adaptive utterances against speech recognition errors and different user vocabulary levels on the MultiWOZ dataset.
When a natural language generation (NLG) component is implemented in a real-world task-oriented dialogue system, it is necessary to generate not only natural utterances as learned on training data but also utterances adapted to the dialogue environment (e.g., noise from environmental sounds) and the user (e.g., users with low levels of understanding ability). Inspired by recent advances in reinforcement learning (RL) for language generation tasks, we propose ANTOR, a method for Adaptive Natural language generation for Task-Oriented dialogue via Reinforcement learning. In ANTOR, a natural language understanding (NLU) module, which corresponds to the user's understanding of system utterances, is incorporated into the objective function of RL. If the NLG's intentions are correctly conveyed to the NLU, which understands a system's utterances, the NLG is given a positive reward. We conducted experiments on the MultiWOZ dataset, and we confirmed that ANTOR could generate adaptive utterances against speech recognition errors and the different vocabulary levels of users.