Project C03: Interpretable machine learning: Explaining Change
Today, machine learning is commonly used in dynamic environments such as social networks, logistics, transportation, retail, finance, and healthcare, where new data is continuously being generated. In order to respond to possible changes in the underlying processes and to ensure that the models that have been learned continue to function reliably, they must be adapted on a continuous basis. These changes, like the model itself, should be kept transparent by providing clear explanations for users. For this, application-specific needs must be taken into account. The researchers working on Project C03 are considering how and why different types of models change from a theoretical-mathematical perspective. Their goal is to develop algorithms that efficiently and reliably detect changes in models and provide intuitive explanations to users.
Research areas: Computer science
Publications
Agnostic Explanation of Model Change based on Feature Importance
M. Muschalik, F. Fumagalli, B. Hammer, E. Huellermeier, KI - Künstliche Intelligenz 36 (2022) 211–224.
“I do not know! but why?” — Local model-agnostic example-based explanations of reject
A. Artelt, R. Visser, B. Hammer, Neurocomputing 558 (2023).
Incremental permutation feature importance (iPFI): towards online explanations on data streams
F. Fumagalli, M. Muschalik, E. Hüllermeier, B. Hammer, Machine Learning 112 (2023) 4863–4903.
iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios
M. Muschalik, F. Fumagalli, R. Jagtani, B. Hammer, E. Huellermeier, in: Proceedings of the World Conference on Explainable Artificial Intelligence (XAI), 2023.
iSAGE: An Incremental Version of SAGE for Online Explanation on Data Streams
M. Muschalik, F. Fumagalli, B. Hammer, E. Huellermeier, in: Machine Learning and Knowledge Discovery in Databases: Research Track - European Conference (ECML PKDD), Springer Nature Switzerland, 2023.
On Feature Removal for Explainability in Dynamic Environments
F. Fumagalli, M. Muschalik, E. Hüllermeier, B. Hammer, in: Proceedings of the European Symposium on Artificial Neural Networks (ESANN), 2023.
SHAP-IQ: Unified Approximation of any-order Shapley Interactions
F. Fumagalli, M. Muschalik, P. Kolpaczki, E. Hüllermeier, B. Hammer, in: Advances in Neural Information Processing Systems (NeurIPS), 2023, pp. 11515--11551.
No learning rates needed: Introducing SALSA - Stable Armijo Line Search Adaptation
P. Kenneweg, T. Kenneweg, F. Fumagalli, B. Hammer, in: 2024 International Joint Conference on Neural Networks (IJCNN), 2024, pp. 1–8.
Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles
M. Muschalik, F. Fumagalli, B. Hammer, E. Huellermeier, in: Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2024, pp. 14388–14396.
SVARM-IQ: Efficient Approximation of Any-order Shapley Interactions through Stratification
P. Kolpaczki, M. Muschalik, F. Fumagalli, B. Hammer, E. Huellermeier, in: Proceedings of The 27th International Conference on Artificial Intelligence and Statistics (AISTATS), PMLR, 2024, pp. 3520–3528.
KernelSHAP-IQ: Weighted Least Square Optimization for Shapley Interactions
F. Fumagalli, M. Muschalik, P. Kolpaczki, E. Hüllermeier, B. Hammer, in: Proceedings of the 41st International Conference on Machine Learning (ICML), PMLR, 2024, pp. 14308–14342.
shapiq: Shapley interactions for machine learning
M. Muschalik, H. Baniecki, F. Fumagalli, P. Kolpaczki, B. Hammer, E. Huellermeier, in: Advances in Neural Information Processing Systems (NeurIPS), 2024, pp. 130324–130357.
Approximating the shapley value without marginal contributions
P. Kolpaczki, V. Bengs, M. Muschalik, E. Hüllermeier, in: Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2024, pp. 13246–13255.
Exact Computation of Any-Order Shapley Interactions for Graph Neural Networks
M. Muschalik, F. Fumagalli, P. Frazzetto, J. Strotherm, L. Hermes, A. Sperduti, E. Hüllermeier, B. Hammer, in: The Thirteenth International Conference on Learning Representations (ICLR), 2025.
Explaining Outliers using Isolation Forest and Shapley Interactions
R. Visser, F. Fumagalli, E. Hüllermeier, B. Hammer, in: Proceedings of the European Symposium on Artificial Neural Networks (ESANN), 2025.
Unifying Feature-Based Explanations with Functional ANOVA and Cooperative Game Theory
F. Fumagalli, M. Muschalik, E. Hüllermeier, B. Hammer, J. Herbinger, in: Proceedings of The 28th International Conference on Artificial Intelligence and Statistics (AISTATS), PMLR, 2025, pp. 5140–5148.
Adaptive Prompting: Ad-hoc Prompt Composition for Social Bias Detection
M. Spliethöver, T. Knebler, F. Fumagalli, M. Muschalik, B. Hammer, E. Hüllermeier, H. Wachsmuth, in: L. Chiruzzo, A. Ritter, L. Wang (Eds.), Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), Association for Computational Linguistics, Albuquerque, New Mexico, 2025, pp. 2421–2449.
Show all publications
