Best Practices for Data-Efficient Modeling in NLG: How to Train Production-Ready Neural Models with Less Data

Abstract

Natural language generation (NLG) is a critical component in conversational systems, owing to its role of formulating a correct and natural text response. Traditionally, NLG components have been deployed using template-based solutions. Although neural network solutions recently developed in the research community have been shown to provide several benefits, deployment of such model-based solutions has been challenging due to high latency, correctness issues, and highdata needs. In this paper, we present approaches that have helped us deploy data-efficient neural solutions for NLG in conversational systems to production. We describe a family of sampling andmodeling techniques to attain production quality with light-weight neural network models usingonly a fraction of the data that would be necessary otherwise, and show a thorough comparison between each. Our results show that domain complexity dictates the appropriate approach toachieve high data efficiency. Finally, we distill the lessons from our experimental findings into alist of best practices for production-level NLG model development, and present them in a brief runbook. Importantly, the end products of all of the techniques are small sequence-to-sequencemodels (~2Mb) that we can reliably deploy in production.