# WikiVideo: Article Generation from Multiple Videos

Source: [WikiVideo: Article Generation from Multiple Videos](<https://www.sota2.com/research/paper/68fa7c6f1e209131cea8e0ef>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-28T10:40:15.178Z

Authors: Alexander Martin, Reno Kriz, William Gantt Walden, Kate Sanders, Hannah Recknor, Eugene Yang, Francis Ferraro, Benjamin Van Durme

Published: 2025-04-01T16:22:15.000Z

## Abstract

We introduce the task of grounded article generation with the goal of creating a Wikipedia-style article from multiple diverse videos about real-world events -- from natural disasters to political elections -- where all the information in the article is supported by video evidence. Videos are intuitive sources for retrieval-augmented generation (RAG), but most contemporary RAG workflows focus heavily on text while existing methods for video-based summarization focus on low-level scene understanding rather than high-level event semantics. To close this gap, we introduce WikiVideo, a benchmark consisting of expert-written articles and densely annotated videos that provide evidence for articles' claims, facilitating the integration of video into RAG pipelines and enabling the creation of in-depth content that is grounded in multimodal sources. We further propose Collaborative Article Generation (CAG), a novel interactive method for article creation from multiple videos. CAG leverages an iterative interaction between an r1-style reasoning model and a VideoLLM to draw higher-level inferences about the target event than is possible with VideoLLMs alone, which fixate on low-level visual features. We benchmark state-of-the-art VideoLLMs and CAG in both oracle retrieval and RAG settings and find that CAG consistently outperforms alternative methods, while suggesting intriguing avenues for future work.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Information Generation on MultiVENT 2.0](<https://www.sota2.com/research/sota/information-generation-on-multivent-2-0.md>): Information Generation; MULTIVENT 2.0; Precision: 65.1.

- [Temporal Grounding on MultiVENT 2.0](<https://www.sota2.com/research/sota/temporal-grounding-on-multivent-2-0.md>): Temporal Grounding; MULTIVENT 2.0; Precision: 51.

- [Citation Quality Evaluation on WikiVideo 1.0 (test)](<https://www.sota2.com/research/sota/citation-quality-evaluation-on-wikivideo-1-0-test.md>): Citation Quality Evaluation; WikiVideo 1.0 (test); CITEP R: 0.

- [Multi-video Grounding and Retrieval on MAGMaR Oracle Track 2026 (val)](<https://www.sota2.com/research/sota/multi-video-grounding-and-retrieval-on-magmar-oracle-track-2026-val.md>): Multi-video Grounding and Retrieval; MAGMaR Oracle Track 2026 (val); Human Evaluation Score: 3.088.

- [Article Generation on MAGMaR oracle (leaderboard snapshot)](<https://www.sota2.com/research/sota/article-generation-on-magmar-oracle-leaderboard-snapshot.md>): Article Generation; MAGMaR oracle (leaderboard snapshot); Human Preference Score: 3.09.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2504.00939>)

[PDF](<https://arxiv.org/pdf/2504.00939.pdf>)
