Prof. Dr. Fabian Fumagalli

About

I am an interim professor at the Chair of Statistical Learning and Data Science.

My research advances explainable AI (XAI) by developing theoretically grounded, efficient methods for interpreting machine learning models, focusing on scalable and reliable Shapley-based explanations.

I co-founded the shapiq package, which extends shap to efficiently handle any-order feature interactions, providing a scalable and flexible toolkit for advanced model interpretation accessible to both researchers and practitioners.

Short Bio

Since October 2025, Fabian Fumagalli is an interim professor for Statistical Learning and Data Science funded by the Munich Center for Machine Learning (MCML). He is a member of the Chair of Statistical Learning and Data Science (Department of Statistics, LMU Munich), which is headed by Prof. Dr. Bernd Bischl. Previously, he was a PostDoc at the HammerLab machine learning group at Bielefeld University, where he finished his PhD under the supervision of Prof. Dr. Barbara Hammer focusing on explainable artificial intelligence (XAI). Before his PhD, he worked as a data science consultant. Fabian Fumagalli holds a Master's degree in Mathematics from TU Berlin and a Bachelor's degree in Mathematics from Freie Universität Berlin.

Contact

Institut für Statistik

Ludwig-Maximilians-Universität München

Room 140

Ludwigstraße 33

D-80539 München

f.fumagalli [at] lmu.de

Teaching

Winter Term 2025/2056
- Seminar (MA): Explaining Black-Box Predictions: A Game-Theoretic Approach to Reliable Model Interpretability)
- Supervised Learning

Research Interests

My main research interest lies in the field of explainable AI (XAI):

Computation of Shapley Value
Computation of Shapley Interactions
Feature-Based Explanations
Cooperative Game Theory in Machine Learning

You Can Find me on

Projects

shapiq: An extension of the shap package for feature interactions. See blog post.

Publications

Baniecki H, Muschalik M, Fumagalli F, Hammer B, Hüllermeier E, Biecek P (2025) Explaining Similarity in Vision-Language Encoders with Weighted Banzhaf Interactions. CoRR abs/2508.05430.
Link | PDF.
Wever M, Muschalik M, Fumagalli F, Lindauer M (2025) HyperSHAP: Shapley Values and Interactions for Hyperparameter Importance. CoRR abs/2502.01276.
Link | PDF.
Muschalik M, Fumagalli F, Frazzetto P, Strotherm J, Hermes L, Sperduti A, Hüllermeier E, Hammer B (2025) Exact Computation of Any-Order Shapley Interactions for Graph Neural Networks Proceedings of the Conference on Learning Representations (ICLR),
Link | PDF.
Fumagalli F, Muschalik M, Hüllermeier E, Hammer B, Herbinger J (2025) Unifying Feature-Based Explanations with Functional ANOVA and Cooperative Game Theory Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), pp. 5140–5148.
Link | PDF.
Spliethöver M, Knebler T, Fumagalli F, Muschalik M, Hammer B, Hüllermeier E, Wachsmuth H (2025) Adaptive Prompting: Ad-hoc Prompt Composition for Social Bias Detection Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL), pp. 2421–2449.
Link | PDF.
Visser R, Fumagalli F, Hüllermeier E, Hammer B (2025) Explaining Outliers using Isolation Forest and Shapley Interactions Proceedings of the European Symposium on Artificial Neural Networks (ESANN),
Link | PDF.
Fumagalli F, Muschalik M, Kolpaczki P, Hüllermeier E, Hammer B (2024) KernelSHAP-IQ: Weighted Least Square Optimization for Shapley Interactions Proceedings of the International Conference on Machine Learning (ICML), pp. 14308–14342.
Link | PDF.
Kolpaczki P, Muschalik M, Fumagalli F, Hammer B, Hüllermeier E (2024) SVARM-IQ: Efficient Approximation of Any-order Shapley Interactions through Stratification Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), pp. 3520–3528.
Link | PDF.
Muschalik M, Fumagalli F, Hammer B, Hüllermeier E (2024) Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles Proceeedings of the AAAI Conference on Artificial Intelligence (AAAI), pp. 14388–14396.
Link | PDF.
Muschalik M, Baniecki H, Fumagalli F, Kolpaczki P, Hammer B, Hüllermeier E (2024) shapiq: Shapley Interactions for Machine Learning Proceedings of Advances in Neural Information Processing Systems (NeurIPS), pp. 130324–130357.
Link | PDF.
Kenneweg P, Kenneweg T, Fumagalli F, Hammer B (2024) No learning rates needed: Introducing SALSA - Stable Armijo Line Search Adaptation Proceedings of the International Joint Conference on Neural Networks (IJCNN),
Link | PDF.
Fumagalli F, Muschalik M, Kolpaczki P, Hüllermeier E, Hammer B (2023) SHAP-IQ: Unified Approximation of any-order Shapley Interactions Proceedings of Advances in Neural Information Processing Systems (NeurIPS), pp. 11515–11551.
Link | PDF.
Fumagalli F, Muschalik M, Hüllermeier E, Hammer B (2023) Incremental permutation feature importance (iPFI): towards online explanations on data streams. Machine Learning 112, 4863–4903.
Link | PDF.
Muschalik M, Fumagalli F, Hammer B, Hüllermeier E (2023) iSAGE: An Incremental Version of SAGE for Online Explanation on Data Streams Machine Learning and Knowledge Discovery in Databases: Research Track - European Conference (ECML PKDD), pp. 428–445.
Link | PDF.
Fumagalli F, Muschalik M, Hüllermeier E, Hammer B (2023) On Feature Removal for Explainability in Dynamic Environments Proceedings of the European Symposium on Artificial Neural Networks (ESANN),
Link | PDF.
Muschalik M, Fumagalli F, Jagtani R, Hammer B, Hüllermeier E (2023) iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios Proceeedings of World Conference on Explainable Artificial Intelligence (xAI),
Link | PDF.
Muschalik M, Fumagalli F, Hammer B, Hüllermeier E (2022) Agnostic Explanation of Model Change based on Feature Importance. Künstliche Intelligenz 36, 211–224.
Link | PDF.