# Improving Truthfulness of Headline Generation

arXiv: https://arxiv.org/abs/2005.00882 · HTML: https://mlsanity.com/paper/2005.00882/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 May 2020.

## Built on by (most cited first, 20 of 20)

- [The Factual Inconsistency Problem in Abstractive Text Summarization: A Survey](https://mlsanity.com/paper/2104.14839/) (2021)
- [Faithful or Extractive? On Mitigating the Faithfulness-Abstractiveness Trade-off in Abstractive Summarization](https://mlsanity.com/paper/2108.13684/) (2021)
- [What's in a Summary? Laying the Groundwork for Advances in Hospital-Course Summarization](https://mlsanity.com/paper/2105.00816/) (2021)
- [Learning to Revise References for Faithful Summarization](https://mlsanity.com/paper/2204.10290/) (2022)
- [Don't Say What You Don't Know: Improving the Consistency of Abstractive Summarization by Constraining Beam Search](https://mlsanity.com/paper/2203.08436/) (2022)
- [Zero-shot Faithfulness Evaluation for Text Summarization with Foundation Language Model](https://mlsanity.com/paper/2310.11648/) (2023)
- [Multi-News+: Cost-efficient Dataset Cleansing via LLM-based Data Annotation](https://mlsanity.com/paper/2404.09682/) (2024)
- [CaPE: Contrastive Parameter Ensembling for Reducing Hallucination in Abstractive Summarization](https://mlsanity.com/paper/2110.07166/) (2021)
- [Questioning the Validity of Summarization Datasets and Improving Their Factual Consistency](https://mlsanity.com/paper/2210.17378/) (2022)
- [Expository Text Generation: Imitate, Retrieve, Paraphrase](https://mlsanity.com/paper/2305.03276/) (2023)
- [Incorporating Question Answering-Based Signals into Abstractive Summarization via Salient Span Selection](https://mlsanity.com/paper/2111.07935/) (2021)
- [Multi-Perspective Abstractive Answer Summarization](https://mlsanity.com/paper/2104.08536/) (2021)
- [HonestBait: Forward References for Attractive but Faithful Headline Generation](https://mlsanity.com/paper/2306.14828/) (2023)
- [Extract, Denoise and Enforce: Evaluating and Improving Concept Preservation for Text-to-Text Generation](https://mlsanity.com/paper/2104.08724/) (2021)
- [Automatic Logical Forms improve fidelity in Table-to-Text generation](https://mlsanity.com/paper/2310.17279/) (2023)
- [Fact-Preserved Personalized News Headline Generation](https://mlsanity.com/paper/2501.11828/) (2025)
- [Toward Improving Coherence and Diversity of Slogan Generation](https://mlsanity.com/paper/2102.05924/) (2021)
- [Leveraging Key Information Modeling to Improve Less-Data Constrained News Headline Generation via Duality Fine-Tuning](https://mlsanity.com/paper/2210.04473/) (2022)
- [Automatic Generation of Factual News Headlines in Finnish](https://mlsanity.com/paper/2212.02170/) (2022)
- [NumHG: A Dataset for Number-Focused Headline Generation](https://mlsanity.com/paper/2309.01455/) (2023)

## Abstract (arXiv)

Most studies on abstractive summarization report ROUGE scores between system and reference summaries. However, we have a concern about the truthfulness of generated summaries: whether all facts of a generated summary are mentioned in the source text. This paper explores improving the truthfulness in headline generation on two popular datasets. Analyzing headlines generated by the state-of-the-art encoder-decoder model, we show that the model sometimes generates untruthful headlines. We conjecture that one of the reasons lies in untruthful supervision data used for training the model. In order to quantify the truthfulness of article-headline pairs, we consider the textual entailment of whether an article entails its headline. After confirming quite a few untruthful instances in the datasets, this study hypothesizes that removing untruthful instances from the supervision data may remedy the problem of the untruthful behaviors of the model. Building a binary classifier that predicts an entailment relation between an article and its headline, we filter out untruthful instances from the supervision data. Experimental results demonstrate that the headline generation model trained on filtered supervision data shows no clear difference in ROUGE scores but remarkable improvements in automatic and manual evaluations of the generated headlines.