FAQ

Who are you guys?

We are a group of passionate undergraduates at the Massachusetts Institute of Technology (MIT) trying to help people understand the news that they read. We are a part of the Tegmark Group under Professor Max Tegmark who worked throughout the summer of 2020 to use a machine learning and natural language processing (NLP) to automatically classify news articles across a wide range of topics.

What data were these trained on?

All of the models in this package were trained on the New York Times Annotated Corpus, which contains over 1.8 million articles manually tagged and/or verified by New York Times editors. After data cleaning, over 1.4 million articles and 538 topic tags were used to train our models.

What is Improve The News, and how I can use it?

This is a news aggregator site developed by a group of researchers at MIT and elsewhere to give you control of your news consumption. Many news sites treat you and your attention as the product they sell to their advertisers, but this one is ad-free and eliminates cost through full automation.

Just as it’s healthier to choose what you eat deliberately than impulsively, it’s more empowering to chose your news diet deliberately on this site than to randomly read what marketers and algorithms elsewhere predict that you’ll impulsively click on. You can make these deliberate choices about topic, bias, style, etc. by adjusting sliders described here.