iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios
Muschalik M, Fumagalli F, Jagtani R, Hammer B, Hüllermeier E (2023)
In: Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I. Longo L (Ed); Communications in Computer and Information Science. Cham: Springer Nature Switzerland: 177-194.
Sammelwerksbeitrag
| Veröffentlicht | Englisch
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Autor*in
Muschalik, Maximilian;
Fumagalli, FabianUniBi ;
Jagtani, Rohit;
Hammer, BarbaraUniBi ;
Hüllermeier, Eyke
Herausgeber*in
Longo, Luca
Einrichtung
Technische Fakultät > AG Machine Learning
SFB/Transregio 318 Constructing Explainability > Projektbereich C: Darstellung und Berechnung von Erklärungen > Teilprojekt C03: Interpretierbares maschinelles Lernen: Erklärbarkeit in dynamischen Umgebungen
Center of Excellence - Cognitive Interaction Technology CITEC > Machine Learning
SFB/Transregio 318 Constructing Explainability > Projektbereich C: Darstellung und Berechnung von Erklärungen > Teilprojekt C03: Interpretierbares maschinelles Lernen: Erklärbarkeit in dynamischen Umgebungen
Center of Excellence - Cognitive Interaction Technology CITEC > Machine Learning
Projekt
Abstract / Bemerkung
Post-hoc explanation techniques such as the well-established partial dependence plot (PDP), which investigates feature dependencies, are used in explainable artificial intelligence (XAI) to understand black-box machine learning models. While many real-world applications require dynamic models that constantly adapt over time and react to changes in the underlying distribution, XAI, so far, has primarily considered static learning environments, where models are trained in a batch mode and remain unchanged. We thus propose a novel model-agnostic XAI framework called incremental PDP (iPDP) that extends on the PDP to extract time-dependent feature effects in non-stationary learning environments. We formally analyze iPDP and show that it approximates a time-dependent variant of the PDP that properly reacts to real and virtual concept drift. The time-sensitivity of iPDP is controlled by a single smoothing parameter, which directly corresponds to the variance and the approximation error of iPDP in a static learning environment. We illustrate the efficacy of iPDP by showcasing an example application for drift detection and conducting multiple experiments on real-world and synthetic data sets and streams.
Erscheinungsjahr
2023
Buchtitel
Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I
Serientitel
Communications in Computer and Information Science
Seite(n)
177-194
ISBN
978-3-031-44063-2
eISBN
978-3-031-44064-9
ISSN
1865-0929
eISSN
1865-0937
Page URI
https://pub.uni-bielefeld.de/record/2983942
Zitieren
Muschalik M, Fumagalli F, Jagtani R, Hammer B, Hüllermeier E. iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios. In: Longo L, ed. Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I. Communications in Computer and Information Science. Cham: Springer Nature Switzerland; 2023: 177-194.
Muschalik, M., Fumagalli, F., Jagtani, R., Hammer, B., & Hüllermeier, E. (2023). iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios. In L. Longo (Ed.), Communications in Computer and Information Science. Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I (pp. 177-194). Cham: Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-44064-9_11
Muschalik, Maximilian, Fumagalli, Fabian, Jagtani, Rohit, Hammer, Barbara, and Hüllermeier, Eyke. 2023. “iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios”. In Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I, ed. Luca Longo, 177-194. Communications in Computer and Information Science. Cham: Springer Nature Switzerland.
Muschalik, M., Fumagalli, F., Jagtani, R., Hammer, B., and Hüllermeier, E. (2023). “iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios” in Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I, Longo, L. ed. Communications in Computer and Information Science (Cham: Springer Nature Switzerland), 177-194.
Muschalik, M., et al., 2023. iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios. In L. Longo, ed. Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I. Communications in Computer and Information Science. Cham: Springer Nature Switzerland, pp. 177-194.
M. Muschalik, et al., “iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios”, Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I, L. Longo, ed., Communications in Computer and Information Science, Cham: Springer Nature Switzerland, 2023, pp.177-194.
Muschalik, M., Fumagalli, F., Jagtani, R., Hammer, B., Hüllermeier, E.: iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios. In: Longo, L. (ed.) Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I. Communications in Computer and Information Science. p. 177-194. Springer Nature Switzerland, Cham (2023).
Muschalik, Maximilian, Fumagalli, Fabian, Jagtani, Rohit, Hammer, Barbara, and Hüllermeier, Eyke. “iPDP: On Partial Dependence Plots in Dynamic Modeling Scenarios”. Explainable Artificial Intelligence. First World Conference, xAI 2023, Lisbon, Portugal, July 26–28, 2023, Proceedings, Part I. Ed. Luca Longo. Cham: Springer Nature Switzerland, 2023. Communications in Computer and Information Science. 177-194.
Export
Markieren/ Markierung löschen
Markierte Publikationen
Web of Science
Dieser Datensatz im Web of Science®Suchen in