Unsupervised Natural Language Generation with Denoising Autoencoders
This addresses the need for NLG systems in tasks like question answering and dialog systems without requiring labeled data, though it appears incremental as it builds on denoising autoencoders.
The paper tackles the problem of generating text from structured data without supervision, achieving higher performance than supervised approaches in at least one domain using only unlabeled text.
Generating text from structured data is important for various tasks such as question answering and dialog systems. We show that in at least one domain, without any supervision and only based on unlabeled text, we are able to build a Natural Language Generation (NLG) system with higher performance than supervised approaches. In our approach, we interpret the structured data as a corrupt representation of the desired output and use a denoising auto-encoder to reconstruct the sentence. We show how to introduce noise into training examples that do not contain structured data, and that the resulting denoising auto-encoder generalizes to generate correct sentences when given structured data.