Expository Text Generation: Imitate, Retrieve, Paraphrase
Nishant Balepur, Jie Huang, Kevin Chen–Chuan Chang · 2023
Expository documents are vital resources for conveying complex information to readers.Despite their usefulness, writing expository text by hand is a challenging process that requires careful content planning, obtaining facts from multiple sources, and the ability to clearly synthesize these facts.To ease these burdens, we propose the task of expository text generation, which seeks to automatically generate an accurate and stylistically consistent expository text for a topic by intelligently searching a knowledge source.We solve our task by developing IRP, a framework that overcomes the limitations of retrieval-augmented models and iteratively performs content planning, fact retrieval, and rephrasing.Through experiments on three diverse, newly-collected datasets, we show that IRP produces factual and organized expository texts that accurately inform readers. 1 1 Code is available at https://github.com/ nbalepur/expository-text-generation. Ground TruthUniversity of Denver is a private institution that was founded in 1864.It has a total undergraduate enrollment of 5,867 (fall 2021), its setting is city, and the campus size is 125 acres... University of Montana is a public institution that was founded in 1893.It has a total undergraduate enrollment of 7,223 (fall 2021), its setting is urban, and the campus size is 220 acres... RAGUniversity of Denver is a public institution founded in 1891.It has a total of 5,867 students (fall 2021), its location is urban, and the campus covers 120 acres... Our Model (IRP)University of Denver is a private institution founded in 1864.It has a total of 5,867 students (fall 2021), it is located in the city, and the campus covers 126 acres... LLaMA+RetrUniversity of Denver is a private institution founded in 1864.It has a total of 11,482 students (fall 2021), its location is urban, and the campus covers 125 acres...