News Recommendation at Scale

With roughly a thousand news articles published by journalists at DPG Media every day, it is a challenge to bring the right articles to the right readers at the right time. An important role of newspapers is to optimally inform readers, balancing between providing news that interests readers, news that provides alternate views or opinions, and news that editors believe readers have to know. Welcome to the world of news personalization.

Example of our personalized push notifications

Why is it hard?

  • ML modeling: designing a recommender that understands both users and news articles and can match them while considering goals such as relevancy, diversity, and recency.
  • Scale: DPG Media is the largest Dutch/Belgian publisher with thousands of journalists producing high-quality content every day. Our users leave traces of their interactions with our products through an event stream with peak volumes around 100K events per second. A recommender must be fast enough to handle these loads in our production environment.

So, News Recommendation

The core idea

Along with a few details

Cool, but does it work?

Filter bubbles and diversity

What we are working on right now

Check out this excellent blogpost by Jay Alammar to learn more about BERT and MLM.

Conclusion

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
DPG Media

DPG Media

We are the tech team behind the digital products of all DPG Media’s brands and internal apps!