Actief project?
Project ended
Project foto of beeld
Logo project NewsTAPAS
Korte beschrijving

Tools and technology for news content providers to automatically adapt personalised items taking into account the user’s profile and context

NewsTAPAS will develop tools and technology for news content providers to automatically adapt personalised items taking into account the user’s profile and context (device, time-of-day, location, etc.). The goal is to establish a novel editorial flow, allowing to create and break-up news stories into its composing story elements, and to build a content adaptation (“how”) engine on top of existing content recommender systems (“what”). NewsTAPAS will significantly increase user engagement and is a unique Flemish collaboration of a public broadcaster (VRT), a multimedia group (Roularta Media Group), an AI service provider (ML2Grow) and imec research groups (SMIT and IDLab).

In order to address the growing issue of information overload and user attention scarcity, existing recommender systems aim to identify news items that best fit a user’s profile. However, the real challenge is to retain the user’s attention and optimize their experience while they are digesting it. The core innovation of NewsTAPAS is the content adaptation engine, that will automatically adapt the presentation of recommended content items to fit the audience needs and user’s context, in order to increase their engagement. Think of selecting the appropriate template for the right audience, and customizing the headline, picture or video thumbnail.

Roularta Media Group, ML2Grow, imec SMIT, imec IDLab
Related articles

Natural Language Processing for media makers: RNN, LSTM and Transformer models

How can deep learning help media makers create content tailored to the reader's preferences? In NewsTAPAS, a project on the personalisation of news, we discovered the need for strengthening automatic capabilities to better summarize articles. Therefore we started working with Natural Language Processing (NLP) and compared recurrent neural networks (RNN), long short-term memory (LTSM), Transformers and their applications within media. In this article we compile some insights and recommendations.

NewsTAPAS brings the news the way you want it

Imagine if an article would adapt itself based on your preferences. Don't have a lot of time for some in-depth reading? Prefer to directly see the headlines? In today's information overload, we are exploring the technological possibilities to better personalise news, not only in terms of content, but also as a format. In the NewsTAPAS project, we are therefore exploring how we can better inform end users by adapting the presentation of a news article to their preferences and context.

New Google DNI project on news personalisation, led by VRT

A new project led by VRT, named NewsTAPAS, has been approved for funding by Google DNI. The project aims to address the growing need for an automated and personalised presentation of news articles. The project team exists of a unique collaboration of VRT with a multimedia publisher (Roularta), an AI start-up (ML2Grow) and imec research groups (SMIT and IDLab). The project will start on May 1st and will last one and a half year.