Abstract: "Textual data have become widely available for data analysis in digital form. It can be anything from medical journals, newspapers, social media to legal texts. In the past this type of data has been categorized manually or analysed with methods from computational linguistics or the area of natural language processing.

Recently we have seen an increased interest in treating unstructured text as data in statistical or probabilistic machine learning models. In this talk we will delve into the area of topic modelling – how we automatically can find themes in textual data using Bayesian statistical models. 

We will also discuss how we can scale these models to millions of documents as well as how these themes or topics can be used in other statistical models or for inference."