Incremental permutation feature importance (iPFI): towards online explanations on data streams
Fumagalli F, Muschalik M, Hüllermeier E, Hammer B (2023)
Machine Learning .
Zeitschriftenaufsatz
| E-Veröff. vor dem Druck | Englisch
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Autor*in
Einrichtung
SFB/Transregio 318 Constructing Explainability > Projektbereich C: Darstellung und Berechnung von Erklärungen > Teilprojekt C03: Interpretierbares maschinelles Lernen: Erklärbarkeit in dynamischen Umgebungen
SFB/Transregio 318 Constructing Explainability > Projektbereich C: Darstellung und Berechnung von Erklärungen > Teilprojekt C01: Gesundes Misstrauen in und durch Erklärungen
Center for Uncertainty Studies (CeUS)
Center of Excellence - Cognitive Interaction Technology CITEC > Machine Learning
Technische Fakultät > AG Machine Learning
SFB/Transregio 318 Constructing Explainability > Projektbereich C: Darstellung und Berechnung von Erklärungen > Teilprojekt C01: Gesundes Misstrauen in und durch Erklärungen
Center for Uncertainty Studies (CeUS)
Center of Excellence - Cognitive Interaction Technology CITEC > Machine Learning
Technische Fakultät > AG Machine Learning
Projekt
Abstract / Bemerkung
Explainable artificial intelligence has mainly focused on static learning scenarios so far. We are interested in dynamic scenarios where data is sampled progressively, and learning is done in an incremental rather than a batch mode. We seek efficient incremental algorithms for computing feature importance (FI). Permutation feature importance (PFI) is a well-established model-agnostic measure to obtain global FI based on feature marginalization of absent features. We propose an efficient, model-agnostic algorithm called iPFI to estimate this measure incrementally and under dynamic modeling conditions including concept drift. We prove theoretical guarantees on the approximation quality in terms of expectation and variance. To validate our theoretical findings and the efficacy of our approaches in incremental scenarios dealing with streaming data rather than traditional batch settings, we conduct multiple experimental studies on benchmark data with and without concept drift.
Stichworte
Incremental learning;
Explainable artificial intelligence;
Feature;
importance;
Permutation feature importance
Erscheinungsjahr
2023
Zeitschriftentitel
Machine Learning
ISSN
0885-6125
eISSN
1573-0565
Page URI
https://pub.uni-bielefeld.de/record/2983727
Zitieren
Fumagalli F, Muschalik M, Hüllermeier E, Hammer B. Incremental permutation feature importance (iPFI): towards online explanations on data streams. Machine Learning . 2023.
Fumagalli, F., Muschalik, M., Hüllermeier, E., & Hammer, B. (2023). Incremental permutation feature importance (iPFI): towards online explanations on data streams. Machine Learning . https://doi.org/10.1007/s10994-023-06385-y
Fumagalli, Fabian, Muschalik, Maximilian, Hüllermeier, Eyke, and Hammer, Barbara. 2023. “Incremental permutation feature importance (iPFI): towards online explanations on data streams”. Machine Learning .
Fumagalli, F., Muschalik, M., Hüllermeier, E., and Hammer, B. (2023). Incremental permutation feature importance (iPFI): towards online explanations on data streams. Machine Learning .
Fumagalli, F., et al., 2023. Incremental permutation feature importance (iPFI): towards online explanations on data streams. Machine Learning .
F. Fumagalli, et al., “Incremental permutation feature importance (iPFI): towards online explanations on data streams”, Machine Learning , 2023.
Fumagalli, F., Muschalik, M., Hüllermeier, E., Hammer, B.: Incremental permutation feature importance (iPFI): towards online explanations on data streams. Machine Learning . (2023).
Fumagalli, Fabian, Muschalik, Maximilian, Hüllermeier, Eyke, and Hammer, Barbara. “Incremental permutation feature importance (iPFI): towards online explanations on data streams”. Machine Learning (2023).
Export
Markieren/ Markierung löschen
Markierte Publikationen
Web of Science
Dieser Datensatz im Web of Science®Suchen in