Site Map

By putting all of your content into nested categories you can give users and search engines access to everything using a menu.

Linas Baltrunas CV

All my carreer I work in the domain of algorithmic personalization. This involves creating new machine learning algorithms or applying known research in the personalization domain. I obtained my Bachelor degree in Computer Science from Vilnius University. Later I moved to a small mountain city Bozen-Bolzano (FUB). I love mountains and AI, therefore, I stayed at FUB a bit longer to obtain my Ph.D degree under the supervision of Prof. Francesco Ricci. During the Ph.D. studies I introduced and evaluated a methodology supporting the full cycle of building a Context Aware Collaborative Filtering Recommender Systems. It includes: two phases of data acquisition, a well tuned algorithm capable of generating explanations for the recommendations, and a graphical user interface. I created more scalable and more accurate algorithms compared to those previously known in the literature. 

After my graduation I moved to Telefonica Research lab, Barcelona, Spain. Here I was further pushing state of the art in personalisation research.

Currently I am working at Netflix where I use data to optimize entertainment business.

You can find my full cv here (updated 2015-06).

Context Aware Collaborative Filtering

Recommender systems are intelligent tools that help on-line users to tame information overload. Most of the current recommender systems assume that users' preferences do not change quickly and previously recorded observations  about what users like or dislike can help in predicting their future choices. This assumption is valid only to some extent. In fact, users' general interests can be relatively stable, but they can also be influenced by many additional and varying factors, here referred as context. For example, weather could play an important role while choosing between going to a beach or visiting an indoor museum.

In the thesis I have studied how contextual information could be exploited to improve the prediction accuracy of Recommender Systems based on the Collaborative Filtering technique. Within this thesis we analyzed several new methods and techniques to perform context-aware Collaborative Filtering. We have also introduced and investigated a comprehensive development methodology for context-aware system. It consists of a two steps data gathering process and a specialized model-based rating prediction algorithm. We validated this methodology on a mobile recommender that provides real-time context-aware recommendations and can explain such recommendations on the base of  the acquired contextual information. We have shown in an experimental study that our context-aware methods can increase the prediction accuracy compared to some context-free approaches and other state of the art context-aware methods. In addition, we carried out a user study that showed that the proposed techniques can also increase the user satisfaction.