Est. 2018
The TREC News Track features modern search tasks in the news domain. In partnership with The Washington Post, we are developing test collections that support the search needs of news readers and news writers in the current news environment. It's our hope that the track will foster research that establishes a new sense for what "relevance" means for news search.
Given a news story, retrieve other news articles that provide important context or background information. The goal is to help a reader understand or learn more about the story or main issues in the current article using the best possible sources.
Given a news story and a list of entities referenced in that story (and linked to Wikipedia), rank the entities reflecting the entities importance to the story and utility in providing that link to the reader in context. The goal is to provide reader context by automatically linking the entities in the article that are most critical to understanding the story.
The TREC Washington Post Collection can be obtained here
The current version of the track guidelines, v1.2, is now available (7/5/2019). Key dates are now set: Background linking runs are due 8/18, entity ranking runs are due 9/22.
In TREC 2019, active participant discussion is happening on Slack.
We have a question-and-answer forum where you can ask questions and (wait for it) get answers. This is primarily intended for general information; if you're planning to participate, you need to register for TREC and join the Slack channel for active participant information.