Jump to content

Wikimedia Research/Showcase/Archive/2021/10

From mediawiki.org

October 2021

[edit]
Theme
Bridging knowledge gaps

October 27, 2021 Video: YouTube

Automatic approaches to bridge knowledge gaps in Wikimedia projects
By WMF Research Team
In order to advance knowledge equity as part of the Wikimedia Movement’s 2030 strategic direction, the Research team at the Wikimedia Foundation has been conducting research to “Address Knowledge Gaps” as one of its main programs. One core component of this program is to develop technologies to bridge knowledge gaps. In this talk, we give an overview on how we approach this task using tools from Machine Learning in four different contexts: section alignment in content translation, link recommendation in structured editing, image recommendation in multimedia knowledge gaps, and the equity of the recommendations themselves. We will present how these models can assist contributors in addressing knowledge gaps. Finally, we will discuss the impact of these models in applications deployed across Wikimedia projects supporting different Product initiatives at the Wikimedia Foundation.
More information on the individual projects:
* Section alignment: meta:Research:Expanding_Wikipedia_articles_across_languages/Inter_language_approach#Section_Alignment
* Link recommendation: meta:Research:Link_recommendation_model_for_add-a-link_structured_task
* Image recommendation: meta:Research:Recommending_Images_to_Wikipedia_Articles
* Equity in recommendations: meta:Research:Prioritization_of_Wikipedia_Articles/Recommendation
Slide deck:
* Slides on figshare