# Earlier Isn't Always Better: Sub-aspect Analysis on Corpus and System Biases in Summarization

arXiv: https://arxiv.org/abs/1908.11723 · HTML: https://mlsanity.com/paper/1908.11723/

## Key facts

- Cited by 28 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Aug 2019.

## Built on by (most cited first, 28 of 28)

- [Benchmarking Cognitive Biases in Large Language Models as Evaluators](https://mlsanity.com/paper/2309.17012/) (2023)
- [Robustness Gym: Unifying the NLP Evaluation Landscape](https://mlsanity.com/paper/2101.04840/) (2021)
- [Measure and Improve Robustness in NLP Models: A Survey](https://mlsanity.com/paper/2112.08313/) (2021)
- [Virtual Personas for Language Models via an Anthology of Backstories](https://mlsanity.com/paper/2407.06576/) (2024)
- [An Empirical Analysis of Uncertainty in Large Language Model Evaluations](https://mlsanity.com/paper/2502.10709/) (2025)
- [Bias in News Summarization: Measures, Pitfalls and Corpora](https://mlsanity.com/paper/2309.08047/) (2023)
- [How well do you know your summarization datasets?](https://mlsanity.com/paper/2106.11388/) (2021)
- [SummVis: Interactive Visual Analysis of Models, Data, and Evaluation for Text Summarization](https://mlsanity.com/paper/2104.07605/) (2021)
- [Robust Natural Language Processing: Recent Advances, Challenges, and Future Directions](https://mlsanity.com/paper/2201.00768/) (2022)
- [Demoting the Lead Bias in News Summarization via Alternating Adversarial Learning](https://mlsanity.com/paper/2105.14241/) (2021)
- [AnswerSumm: A Manually-Curated Dataset and Pipeline for Answer Summarization](https://mlsanity.com/paper/2111.06474/) (2021)
- [Improving the Robustness of Summarization Systems with Dual Augmentation](https://mlsanity.com/paper/2306.01090/) (2023)
- [How "Multi" is Multi-Document Summarization?](https://mlsanity.com/paper/2210.12688/) (2022)
- [Unveiling the Lack of LVLM Robustness to Fundamental Visual Variations: Why and Path Forward](https://mlsanity.com/paper/2504.16727/) (2025)
- [Multi-Perspective Abstractive Answer Summarization](https://mlsanity.com/paper/2104.08536/) (2021)
- [SumREN: Summarizing Reported Speech about Events in News](https://mlsanity.com/paper/2212.01146/) (2022)
- [Understanding Position Bias Effects on Fairness in Social Multi-Document Summarization](https://mlsanity.com/paper/2405.01790/) (2024)
- [Dynamic Sliding Window for Meeting Summarization](https://mlsanity.com/paper/2108.13629/) (2021)
- [Summarization of Opinionated Political Documents with Varied Perspectives](https://mlsanity.com/paper/2411.04093/) (2024)
- [LaMSUM: Amplifying Voices Against Harassment through LLM Guided Extractive Summarization of User Incident Reports](https://mlsanity.com/paper/2406.15809/) (2024)
- [Reranking-based Generation for Unbiased Perspective Summarization](https://mlsanity.com/paper/2506.15925/) (2025)
- [Analysis of GraphSum's Attention Weights to Improve the Explainability of Multi-Document Summarization](https://mlsanity.com/paper/2105.11908/) (2021)
- [Input Order Shapes LLM Semantic Alignment in Multi-Document Summarization](https://mlsanity.com/paper/2512.02665/) (2025)
- [Sample Efficient Multimodal Semantic Augmentation for Incremental Summarization](https://mlsanity.com/paper/2303.04361/) (2023)
- [Evaluating Robustness of Dialogue Summarization Models in the Presence of Naturally Occurring Variations](https://mlsanity.com/paper/2311.08705/) (2023)
- [Exploiting Representation Bias for Data Distillation in Abstractive Text Summarization](https://mlsanity.com/paper/2312.06022/) (2023)
- [AspirinSum: an Aspect-based utility-preserved de-identification Summarization framework](https://mlsanity.com/paper/2406.13947/) (2024)
- [Fair Summarization: Bridging Quality and Diversity in Extractive Summaries](https://mlsanity.com/paper/2411.07521/) (2024)

## Abstract (arXiv)

Despite the recent developments on neural summarization systems, the underlying logic behind the improvements from the systems and its corpus-dependency remains largely unexplored. Position of sentences in the original text, for example, is a well known bias for news summarization. Following in the spirit of the claim that summarization is a combination of sub-functions, we define three sub-aspects of summarization: position, importance, and diversity and conduct an extensive analysis of the biases of each sub-aspect with respect to the domain of nine different summarization corpora (e.g., news, academic papers, meeting minutes, movie script, books, posts). We find that while position exhibits substantial bias in news articles, this is not the case, for example, with academic papers and meeting minutes. Furthermore, our empirical study shows that different types of summarization systems (e.g., neural-based) are composed of different degrees of the sub-aspects. Our study provides useful lessons regarding consideration of underlying sub-aspects when collecting a new summarization dataset or developing a new system.