Semi-Supervised Neural Text Generation by Joint Learning of Natural Language Generation and Natural Language Understanding Models

TL;DR

This paper introduces a semi-supervised deep learning approach for natural language generation that jointly trains NLG and NLU models, enabling effective learning from limited annotated data and abundant unannotated data.

Contribution

It proposes a novel joint learning framework for NLG and NLU models that reduces dependency on annotated datasets for end-to-end natural language generation.

Findings

Achieves competitive results with limited annotated data

Effectively leverages non-annotated datasets for training

Does not require pre-processing or re-scoring tricks

Abstract

In Natural Language Generation (NLG), End-to-End (E2E) systems trained through deep learning have recently gained a strong interest. Such deep models need a large amount of carefully annotated data to reach satisfactory performance. However, acquiring such datasets for every new NLG application is a tedious and time-consuming task. In this paper, we propose a semi-supervised deep learning scheme that can learn from non-annotated data and annotated data when available. It uses an NLG and a Natural Language Understanding (NLU) sequence-to-sequence models which are learned jointly to compensate for the lack of annotation. Experiments on two benchmark datasets show that, with limited amount of annotated data, the method can achieve very competitive results while not using any pre-processing or re-scoring tricks. These findings open the way to the exploitation of non-annotated datasets which is the current bottleneck for the E2E NLG system development to new applications.