Neural Text Generation from Structured and Unstructured Data

Abstract

A number of researchers have recently questioned the necessity of increasingly complex neural network (NN) architectures. In particular, several recent papers have shown that simpler, properly tuned models are at least competitive across several natural language processing tasks. In this thesis, we show that this is also the case for text generation from structured and unstructured data. Specifically, we consider neural table-to-text generation and neural question generation (NQG) tasks for text generation from structured and unstructured data respectively. Table-to-text generation aims to generate a description based on a given table, and NQG is the task of generating a question from a given passage where the generated question can be answered by a certain sub-span of the passage using NN models. Experiments demonstrate that a basic attention-based sequence-to-sequence model trained with exponential moving average technique achieves state of the art in both tasks. We further investigate using reinforcement learning with different reward functions to refine our pre-trained model for both tasks.