Skip to content

Research

My research lies at the intersection of computational mathematics and reliable machine learning, with a specific focus on Conformal Prediction, decision-making under uncertainty, and non-parametric statistics.

Peer-Reviewed Publications

  • A Fast, Closed-Form Bandwidth Selector for the Beta Kernel Density Estimator J. Hallberg Szabadváry To appear in Journal of Computational and Graphical Statistics (2026) [ arXiv ] | [ GitHub ]

  • Classification with Reject Option: Distribution-free Error Guarantees via Conformal Prediction J. Hallberg Szabadváry, T. Löfström, U. Johansson, C. Sönströd, E. Ahlberg, L. Carlsson Machine Learning with Applications (2025) [ DOI ]

  • Beyond Conformal Predictors: Adaptive Conformal Inference with Confidence Predictors J. Hallberg Szabadváry, T. Löfström Pattern Recognition (2025) [ DOI ]

  • Conformal Predictive Decision Making: A Comparative Study with Bayesian and Point-Predictive Methods S. Lanngren, M. Toremark, J. Hallberg Szabadváry Proceedings of Machine Learning Research, Vol 266 (2025)

  • online-cp: a Python Package for Online Conformal Prediction J. Hallberg Szabadváry, T. Löfström, R. Matela Proceedings of Machine Learning Research, Vol 266 (2025) [ Code ]

  • Conformal LLM Multi-label Text Classification with Binary Relevance Approach V. Ornbratt, J. Hallberg Szabadváry Proceedings of Machine Learning Research, Vol 266 (2025)

  • Adaptive Conformal Inference for Multi-Step Ahead Time-Series Forecasting Online J. Hallberg Szabadváry Proceedings of the 13th Symposium on Conformal and Probabilistic Prediction with Applications (COPA 2024) [ PDF ]

  • Evaluation of conformal-based probabilistic forecasting methods for short-term wind speed forecasting S. Althoff, J. Hallberg Szabadváry, J. Anderson, L. Carlsson Proceedings of COPA 2023, PMLR 204:100-115

  • On Qualitative Analysis of a Discrete Time SIR Model J. Hallberg Szabadváry, Y. Zhou Chaos, Solitons & Fractals: X (2021) [ DOI ]

Book Chapters & Special Contributions

  • Application of Confidence and Probabilistic Models to Practical Problems L. Carlsson, J. Hallberg Szabadváry, E. Ahlberg, J. Gammerman In: The Importance of Being Learnable: Essays Dedicated to Alexander Gammerman (2026) [ DOI ]

Preprints & Submitted Work

  • Conformal Blindness: A Note on A-Cryptic change-points J. Hallberg Szabadváry arXiv:2601.01147 (2026) [ DOI ]

  • Ensured: Explanations for Decreasing the Epistemic Uncertainty in Predictions H. Löfström, T. Löfström, J. Hallberg Szabadváry Preprint (2024) [ arXiv:2410.05479 ]

Software

  • Beta KDE Bandwidth Selector ("HS Rule") Implementations (2025–2026) Implementations of the closed-form bandwidth selector for boundary-corrected kernel density estimation.

    • Python (beta-kde): Scikit-learn compatible library. [ Documentation | GitHub ]
    • Julia (BetaKDE.jl): Native Julia package. [ GitHub ]
    • R (kdensity): Integrated as the default beta kernel bandwidth selector ("HS") in the standard CRAN package. [ CRAN | GitHub ]

  • online-cp (2025) A comprehensive Python package for Online Conformal Prediction. [ GitHub ]

Theses

  • Master Thesis: Single- and Multi-fidelity Bandits in Monte Carlo Tree Search – From the casino to mobile network optimization (2022)
  • Bachelor Thesis: A Fairly Complete Qualitative Analysis of a Discrete SIR Epidemiological Model (2021)