Natural Language Generation as Planning under Uncertainty for Spoken Dialogue Systems

Abstract

AbstractWe present and evaluate a new model for Natural Language Generation (NLG) in Spoken Dialogue Systems, based on statistical planning, given noisy feedback from the current generation context (e.g. a user and a surface realiser). The model is adaptive and incremental at the turn level, and optimises NLG actions with respect to a data-driven objective function. We study its use in a standard NLG problem: how to present information (in this case a set of search results) to users, given the complex trade-offs between utterance length, amount of information conveyed, and cognitive load. We set these trade-offs in an objective function by analysing existing match data. We then train a NLG policy using Reinforcement Learning (RL), which adapts its behaviour to noisy feedback from the current generation context. This policy is compared to several baselines derived from previous work in this area. The learned policy significantly outperforms all the prior approaches.KeywordsReinforcement LearningAdaptivitySpoken Dialogue SystemsInformation PresentationIncremental NLGOptimisationdata-driven methods