Achtung:

Sie haben Javascript deaktiviert!
Sie haben versucht eine Funktion zu nutzen, die nur mit Javascript möglich ist. Um sämtliche Funktionalitäten unserer Internetseite zu nutzen, aktivieren Sie bitte Javascript in Ihrem Browser.

Further information:

Research interests

  • Structured prediction investigates intelligent algorithms for supervised problems with complex data, i.e. data that cannot be easily described as a set of instances with a fixed set of attributes.
  • Factorization models address problems resulting from categorical variables with many levels (e.g. in recommender systems). Interactions between these variables can then be captured in models by interactions between the latent features.
  • Multi-relational learning provides an umbrella for problems with a complex, relational structure. Predictions in dyadic relations may be tackled by matrix factorization models, predictions in higher-order relations by tensor factorization models; multi-relational factorization models can describe problems when additional covariate relations are available.
  • Large scale robust online learning and inference algorithms tackle convex and non-convex problems with data in the order of petabytes.

    Applications

    • Recommender systems are dynamic adaptive systems for personalization that help users to chose between alternatives and to find items they are interested in; they learn user preferences from collective past user behavior and are very popular in environments such as e-commerce, web and IPTV multimedia applications, social networks, and online tutoring systems.
    • Online experimenting / significance testing
    • Near real-time data warehousing

    The University for the Information Society