Bibblio Blog

Thought pieces on media, publishing, machine learning and content recommendation

Building a related article recommender system 6 December 2018 Find out what we learned at the 7th RecSys London. Maya Hristakeva, who works at Elsevier, gave a talk titled: ‘Beyond Collaborative Filtering: Learning to Rank Research Articles’. Here's a detailed recap on how her team built, iterated and improved the Science Direct related article recommender. Reaching for the Sky at our latest recommender systems meetup 4 July 2018 Find out what we learned at our fifth Recsys London meetup. With a Principal Data Scientist from Sky, a Data Science Consultant and Marketeer, and a Senior Research Engineer from Sky all contributing, there were lots of great insights! Our favourite takeaways: food, dating and other recommendations 25 June 2018 Get the insights from Bibblio Events' inaugural Recommender Systems Meetup in Berlin: RecSys Berlin. We learn all about how dating and food sometimes need different approaches - and different ways of doing recommendations too ;) What We Learnt at Our Inaugural Recommender Meetup in NYC 24 May 2018 Find out what we learnt from the presentations at the inaugural RecSys New York City meetup - there were lots of great insights on the right systems to solve problems from recommending news stories, to music and recipe recommendations! Food for thought from London: the future of recommender systems 1 May 2018 Discover what happened at the last get together of the RecSys London meetup - three great talks for anyone interested in algorithm development and recommender systems! (and pizza, courtesy of our friends at Deliveroo) Insights from an evening with recommender systems experts 8 February 2018 Bibblio Labs organised the third RecSys meetup in London, and we heard all about hacking a recommender, evaluation metrics and production tools for recommenders What are the three ways to build a recommender system when you don’t have audience data? 6 February 2018 Learn three ways to build a recommender system when you don't have data on your users or audience from data scientist and algorithm builder Dr. Mahbub Gani What matters most when building algorithms - people, process or code? 1 February 2018 Learn the formula behind Bibblio’s method for building algorithms for its recommender system: harnessing the whole team and valuing process as well as code. The secret lives of algorithms 30 November 2017 Algorithms have a big impact, but their workings are often hidden. This lack of transparency has led to diminishing public trust in them. How can we do better? Making the most of every page 28 November 2017 Article pages can’t just tell great stories anymore. They have to perform the function of a homepage too. Find out how good recommendations help you do this. Taking a leap into the deep learning space 27 October 2017 Head of Bibblio Labs, Robbert van der Pluijm, talks deep learning at the 8th Recommender Systems meetup in Amsterdam, in October 2017. Streaming architecture to the rescue 26 September 2017 Bibblio's Lead Data Scientist Dr Mahbub Gani presents how streaming architecture solved the recommendation company's scale and catalogue updating.