The world is being quietly rearranged by people who write very long documents.


The title they went with ClaimPT: A Portuguese Dataset of Annotated Claims in News Articles Noisy translates that to

First Portuguese news dataset for automated fact-checking


Researchers released ClaimPT, a dataset of 1,308 Portuguese news articles with 6,875 manually labeled claims — the first large, freely available resource for training automated fact-checking systems in Portuguese. This makes it possible for researchers and companies outside English-speaking countries to build tools that can automatically spot false claims in news, rather than relying entirely on manual checking.
Fact-checking is slower than misinformation spreads, and automating it requires training data that currently exists almost only in English; releasing Portuguese training data removes one structural barrier to deploying fact-checking tools in a language spoken by 250+ million people.

If you insist
Read the original →