# Fact-based Text Editing

arXiv: https://arxiv.org/abs/2007.00916 · HTML: https://mlsanity.com/paper/2007.00916/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Jul 2020.

## Built on by (most cited first, 14 of 14)

- [RARR: Researching and Revising What Language Models Say, Using Language Models](https://mlsanity.com/paper/2210.08726/) (2022)
- [SummIt: Iterative Text Summarization via ChatGPT](https://mlsanity.com/paper/2305.14835/) (2023)
- [PURR: Efficiently Editing Language Model Hallucinations by Denoising Language Model Corruptions](https://mlsanity.com/paper/2305.14908/) (2023)
- [Learning Structural Edits via Incremental Tree Transformations](https://mlsanity.com/paper/2101.12087/) (2021)
- [AmbigNLG: Addressing Task Ambiguity in Instruction for NLG](https://mlsanity.com/paper/2402.17717/) (2024)
- [Converge to the Truth: Factual Error Correction via Iterative Constrained Editing](https://mlsanity.com/paper/2211.12130/) (2022)
- [XATU: A Fine-grained Instruction-based Benchmark for Explainable Text Updates](https://mlsanity.com/paper/2309.11063/) (2023)
- [Improving Factual Error Correction by Learning to Inject Factual Errors](https://mlsanity.com/paper/2312.07049/) (2023)
- [A Systematic Review of Data-to-Text NLG](https://mlsanity.com/paper/2402.08496/) (2024)
- [HyperEdit: Unlocking Instruction-based Text Editing in LLMs via Hypernetworks](https://mlsanity.com/paper/2512.12544/) (2025)
- [Medico: Towards Hallucination Detection and Correction with Multi-source Evidence Fusion](https://mlsanity.com/paper/2410.10408/) (2024)
- [POLygraph: Polish Fake News Dataset](https://mlsanity.com/paper/2407.01393/) (2024)
- [Draft, Command, and Edit: Controllable Text Editing in E-Commerce](https://mlsanity.com/paper/2208.05623/) (2022)
- [WikiIns: A High-Quality Dataset for Controlled Text Editing by Natural Language Instruction](https://mlsanity.com/paper/2310.05009/) (2023)

## Abstract (arXiv)

We propose a novel text editing task, referred to as textitfact-based text editing, in which the goal is to revise a given document to better describe the facts in a knowledge base (e.g., several triples). The task is important in practice because reflecting the truth is a common requirement in text editing. First, we propose a method for automatically generating a dataset for research on fact-based text editing, where each instance consists of a draft text, a revised text, and several facts represented in triples. We apply the method into two public table-to-text datasets, obtaining two new datasets consisting of 233k and 37k instances, respectively. Next, we propose a new neural network architecture for fact-based text editing, called textscFactEditor, which edits a draft text by referring to given facts using a buffer, a stream, and a memory. A straightforward approach to address the problem would be to employ an encoder-decoder model. Our experimental results on the two datasets show that textscFactEditor outperforms the encoder-decoder approach in terms of fidelity and fluency. The results also show that textscFactEditor conducts inference faster than the encoder-decoder approach.