Julian Rodemann

Statistics, Ludwig-Maximilians-University Munich

Artificial intelligence (AI) is already taking a lot of work off our hands.

But can we rely on AI? Major breakthroughs in AI research are based on machine learning using huge amounts of data – including data that the AI itself has generated or at least influenced. This data sometimes has complex dependency structures. In my PhD, I am trying to develop statistical guarantees for learning from such unequally generated data – with the goal of making AI more reliable and safer.

julian.rodemann@stat.uni-muenchen.de
Profile
Publications

Publications

  • Nalenz, Malte; Rodemann, Julian; Augustin, Thomas (2024)

    Learning de-biased regression trees and forests from complex samples

    Contribution / chapter
    Special Issue of the ACML 2023, Machine Learning, P. 1-20, Springer
    Project: GC-Prom
  • Rodemann, Julian; Goschenhofer, Jann; Dorigatti, Emilio; Nagler, Thomas; Augustin, Thomas (2023)

    Approximately Bayes-Optimal Pseudo-Label Selection

    Conference paper
    Evans, Robin J.; Shpitser, Ilya, Proceedings of Conference on Uncertainty in Artificial Intelligence, Nr. 216, Pittsburgh, Pennslyvania, US
    Project: GC-Prom