oh

Machine learning & statistical inference

Oliver Hennhöfer

I study how machine learning systems can make uncertainty explicit—and what that means for the decisions we make with them.

Research staff at Karlsruhe University of Applied Sciences. My work focuses on conformal inference, anomaly detection, and error control in sequential testing.

Research in practice

nonconform

Conformal anomaly and change-point detection in Python. Moving beyond heuristic thresholds to make uncertainty explicit.

online-fdr

False discovery rate control for sequential hypothesis testing. Python tools for testing hypotheses as they arrive.

aberrant

Online anomaly detection for streaming data. Python methods and tooling for identifying unusual observations as a stream unfolds.

Selected publications

Full record ↗
  1. Between Resolution Collapse and Variance Inflation: Weighted Conformal Anomaly Detection in Low-Data Regimes

    O Hennhöfer, C Preisach
    arXiv preprint arXiv:2603.23205 · 2026

  2. Conformal Anomaly Detection in Python: Moving Beyond Heuristic Thresholds with 'nonconform'

    O Hennhöfer, M Kirsch, C Preisach
    arXiv preprint arXiv:2605.13642 · 2026

Writing

All writing ↗