Natural Language Generation in Dialogue using Lexicalized and Delexicalized Data

Natural language generation plays a critical role in any spoken dialogue system. We present a new approach to natural language generation using recurrent neural networks in an encoder-decoder framework. In contrast with previous work, our model uses both lexicalized and delexicalized versions of slot-value pairs for each dialogue act. This allows our model to learn from all available data, rather than being restricted to learning only from delexicalized slot-value pairs. We show that this helps our model generate more natural sentences with better grammar. We further improve our model's performance by initializing its weights from a pretrained language model. Human evaluation of our best-performing model indicates that it generates sentences which users find more natural and appealing.