During his stay he will focus on researching methods for predicting popularity of online content in social media. The researcher will work at estimating the popularity of visual content, such as images or videos, using social and visual cues. Using machine learning algorithms, such as Support Vector Regression and multivariate regression with Radial Basis Functions, one can determine the future evolution trends in the number of views for images and videos published on social media platforms such as Facebook or Twitter. For the purposes of the prediction, one can use both social cues, such as number of likes, shares, comments, as well as visual cues, such as number of faces in the scene or presence of particular objects. Dr. Trzciński will analyse the differences between those features, measure their importance in the context of popularity prediction and develop new methods for popularity prediction. This information proves to be invaluable in the context of information overflow, since from journalists’ perspective, the application of this know-how can enable optimisation of the visual content towards the needs of the audience.