CDAE | FACTS.lab

Understanding event descriptions is a central aspect of language processing, but current approaches focus overwhelmingly on single sentences or documents. Aggregating information about an event across documents can offer a much richer understanding of what happened, who was involved, and how different sources report on the same underlying facts. The Cross-Document Argument Extraction (CDAE) project develops methods and resources for extracting and synthesizing information about events from multiple source documents.

A key challenge in cross-document event understanding is determining how to align and aggregate information across sources that may describe the same event differently or provide complementary details. This project addresses two main tasks: (i) cross-document argument extraction—identifying the arguments of events (who did what to whom, when, where, and why) across different accounts of the same event; and (ii) cross-document event-keyed summarization—generating coherent summaries of specific events by synthesizing information from multiple sources.

CDAE is supported by a National Science Foundation collaborative grant (BCS-2040831/BCS-2040820) and the Human Language Technology Center of Excellence at Johns Hopkins University.

Website

Project Members

Aaron Steven White

Siddharth Vashishtha

William Gantt Walden

Angela Cao

Claire Sun

Alexander Martin

Weiwei Gu

Chihsheng Jin

Curisia Allen

Papers

Walden, William, Pavlo Kuchmiichuk, Alexander Martin, Chihsheng Jin, Angela Cao, Claire Sun, Curisia Allen & Aaron White. 2025. Cross-Document Event-Keyed Summarization. In Hao Fei, Kewei Tu, Yuhui Zhang, Xiang Hu, Wenjuan Han, Zixia Jia, Zilong Zheng, et al. (eds.), Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025), 218–241. Vienna, Austria: Association for Computational Linguistics.

pdf data + code bib

@inproceedings{walden_cross-document_2025,
    title = "Cross-Document Event-Keyed Summarization",
    author = "Walden, William  and
      Kuchmiichuk, Pavlo  and
      Martin, Alexander  and
      Jin, Chihsheng  and
      Cao, Angela  and
      Sun, Claire  and
      Allen, Curisia  and
      White, Aaron",
    editor = "Fei, Hao  and
      Tu, Kewei  and
      Zhang, Yuhui  and
      Hu, Xiang  and
      Han, Wenjuan  and
      Jia, Zixia  and
      Zheng, Zilong  and
      Cao, Yixin  and
      Zhang, Meishan  and
      Lu, Wei  and
      Siddharth, N.  and
      {\O}vrelid, Lilja  and
      Xue, Nianwen  and
      Zhang, Yue",
    booktitle = "Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025)",
    month = aug,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.xllm-1.19/",
    pages = "218--241",
    ISBN = "979-8-89176-286-2",
    abstract = "Event-keyed summarization (EKS) requires summarizing a specific event described in a document given the document text and an event representation extracted from it. In this work, we extend EKS to the cross-document setting (CDEKS), in which summaries must synthesize information from accounts of the same event as given by multiple sources. We introduce **SEAMuS** (**S**ummaries of **E**vents **A**cross **Mu**ltiple **S**ources), a high-quality dataset for CDEKS based on an expert reannotation of the FAMuS dataset for cross-document argument extraction. We present a suite of baselines on SEAMuS{---}covering both smaller, fine-tuned models, as well as zero- and few-shot prompted LLMs{---}along with detailed ablations and a human evaluation study, showing SEAMuS to be a valuable benchmark for this new task."
}

Copied!

Gantt, William, Alexander Martin, Pavlo Kuchmiichuk & Aaron White. 2024. Event-Keyed Summarization. In Findings of the Association for Computational Linguistics: EMNLP 2024, 7333–7345. Miami, Florida, USA: Association for Computational Linguistics.

pdf data doi bib

@inproceedings{gantt_event_2024,
    title = "Event-Keyed Summarization",
    author = "Gantt, William  and
      Martin, Alexander  and
      Kuchmiichuk, Pavlo  and
      White, Aaron",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024",
    month = nov,
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.findings-emnlp.431",
    doi = "10.18653/v1/2024.findings-emnlp.431",
    pages = "7333--7345",
}

Copied!

Vashishtha, Siddharth, Alexander Martin, William Gantt, Benjamin Van Durme & Aaron White. 2024. FAMuS: Frames Across Multiple Sources. In Kevin Duh, Helena Gomez & Steven Bethard (eds.), Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), 8250–8273. Mexico City, Mexico: Association for Computational Linguistics.

pdf data doi bib

@inproceedings{vashishtha_famus_2024,
  title = "{FAM}u{S}: Frames Across Multiple Sources",
  author = "Vashishtha, Siddharth  and
    Martin, Alexander  and
    Gantt, William  and
    Van Durme, Benjamin  and
    White, Aaron",
  editor = "Duh, Kevin  and
    Gomez, Helena  and
    Bethard, Steven",
  booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)",
  month = jun,
  year = "2024",
  address = "Mexico City, Mexico",
  publisher = "Association for Computational Linguistics",
  url = "https://aclanthology.org/2024.naacl-long.457",
  doi = "10.18653/v1/2024.naacl-long.457",
  pages = "8250--8273",
  abstract = "Understanding event descriptions is a central aspect of language processing, but current approaches focus overwhelmingly on single sentences or documents. Aggregating information about an event across documents can offer a much richer understanding. To this end, we present FAMuS, a new corpus of Wikipedia passages that report on some event, paired with underlying, genre-diverse (non-Wikipedia) source articles for the same event. Events and (cross-sentence) arguments in both report and source are annotated against FrameNet, providing broad coverage of different event types. We present results on two key event understanding tasks enabled by FAMuS: source validation{---}determining whether a document is a valid source for a target report event{---}and cross-document argument extraction{---}full-document argument extraction for a target event from both its report and the correct source article.",
}

Copied!