Two Birds, One Stone: A Simple, Unified Model for Text Generation from Structured and Unstructured Data

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 3864–3870 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 3864 Two Birds, One Stone: A Simple, Unified Model for Text Generation from Structured and Unstructured Data Hamidreza Shahidi,1,2 Ming Li,1,2 and Jimmy Lin1,2 1 David R. Cheriton School of Computer Science, University of Waterloo 2 RSVP.ai {h24shahi, mli, jimmylin}@uwaterloo.ca Abstract A number of researchers have recently ques- tioned the necessity of increasingly complex neural network (NN) architectures. In partic- ular, several recent papers have shown that simpler, properly tuned models are at least competitive across several NLP tasks. In this work, we show that this is also the case for text generation from structured and un- structured data. We consider neural table- to-text generation and neural question gener- ation (NQG) tasks for text generation from structured and unstructured data, respectively. Table-to-text generation aims to generate a de- scription 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. Experimen- tal results demonstrate that a basic attention- based seq2seq model trained with the expo- nential moving average technique achieves the state of the art in both tasks. Code is avail- able at https://github.com/h-shahidi/ 2birds-gen. 1 Introduction Recent NLP literature can be characterized as in- creasingly complex neural network architectures that eke out progressively smaller gains over previ- ous models. Following a previous line of research (Melis et al., 2018; Mohammed et al., 2018; Ad- hikari et al., 2019), we investigate the necessity of such complicated neural architectures. In this work, our focus is on text generation from structured and unstructured data by considering description gener- ation from a table and question generation from a passage and a target answer. More specifically, the goal of the neural table- to-text generation task is to generate biographies based on Wikipedia infoboxes (structured data). An infobox is a factual table with a number of fields Target Output: Sir Bernard Augustus Keen FRS (5 September 1890 – 5 August 1981) was a British soil scientist and Fellow of University College London. Figure 1: An example infobox from the WIKIBIO dataset and the corresponding target output description. Passage: Hydrogen is commonly used in power stations as a coolant in generators due to a num- ber of favorable properties that are a direct result of its light diatomic molecules. Answer: as a coolant in generators Question: How is hydrogen used at power sta- tions? Table 1: A sample (passage, answer, question) triple from the SQuAD dataset. (e.g., name, nationality, and occupation) describing a person. For this task, we use the WIKIBIO dataset (Lebret et al., 2016) as the benchmark dataset. Fig- ure 1 shows an example of a biographic infobox as well as the target output textual description. Automatic question generation aims to gener- ate a syntactically correct, semantically meaning- ful and relevant question from a natural language …

Content truncated to 500 words