MediaFutures partner Bergens Tidende (BT) has been nominated as a finalist for the Global INMA Media Award 2024.
This nomination in the category Best Idea to Encourage Reader Engagement, Regional Brands, is based on a combined submission which was partly due to the WP2 Master’s thesis project “Using content- and behavioural data for recommendations in the Norwegian news market” under supervision of MediaFutures professor Mehdi Elahi and researcher Thomas Husken.
Master student Peter Kolbeinsen Klingenberg investigated in his thesis the viability of using both content- and user-behavioural data to generate recommendations in the Norwegian media market in collaboration with Bergens Tidende (BT). The first technique investigated is collaborative-based filtering, built on the recommender models Alternating Least Squares, Bayesian Personalized Ranking and Logistic Matrix Factorization, which is specially tailored to use user behavioural data as input. The second technique investigated is content-based filtering, built on a state-of-the-art architecture named BERT, specially trained to draw the semantic content of sentences in the Norwegian language.
The results of the collaborative filtering technique have shown increased performance in different stages of filtering, in addition to the importance of hyperparameter tuning. The results have shown promising performance based on the content-based filtering technique, indicating that this attracts users’ interest at a greater scale, even the groups that usually show less interaction.
Based on this thesis, BT introduced “The topic ranker”. A tool that looks at each user’s reading pattern on a section level, and recommended similar stories. They combined this signal with their other rankers, and increased the CTR on stories on the frontpage with 9%.
The ML initiative uses sBERT, and provides similar articles based on the sentiment and content of what you already read.
This model is trained by the AI lab of the national library on a wide variety of Norwegian text – both Bokmål and Nynorsk – from the last 200 years and further finetuned for the purpose of calculating textual similarity. In A/B tests comparing it to our old recommender system, we saw a consistent lift of 20%. We are currently moving this from proof of concept, to a system that can scale.
INMA announced 193 finalists in the 2024 INMA Global Media Awards, highlighting themes around the use of AI in newsroom innovation, promoting the public good campaigns, reaching younger audiences, sports, the war in Ukraine, and user engagement.
First-place winners, regional winners, and the global “Best In Show” will be unveiled April 25 at the Victoria & Albert Museum in London, as the culmination of the INMA World Congress of News Media being held April 22-26.