Beyond Black Boxes: Adaptive XAI for Dynamic Data Pipelines
Otmane Azeroual, Otmane Azeroual
2025
Abstract
The increasing use of real-time data streams in application areas such as the Internet of Things (IoT), financial analytics, and social media demands highly flexible and self-adaptive data pipelines. Modern AI techniques enable the automatic adjustment of these pipelines to dynamically changing data landscapes; however, their decision-making processes often remain opaque and difficult to interpret. This paper presents and evaluates novel approaches for integrating Explainable Artificial Intelligence (XAI) into self-adaptive real-time data pipelines. The goal is to ensure transparent and interpretable data processing while meeting the requirements of real-time capability and scalability. The proposed methods aim to strengthen trust in automated systems and simultaneously address regulatory demands. Initial experimental results demonstrate promising improvements in both explainability and adaptivity without significant performance degradation.
DownloadPaper Citation
in Harvard Style
Azeroual O. (2025). Beyond Black Boxes: Adaptive XAI for Dynamic Data Pipelines. In Proceedings of the 17th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management - Volume 2: KMIS; ISBN 978-989-758-769-6, SciTePress, pages 428-437. DOI: 10.5220/0013736100004000
in Bibtex Style
@conference{kmis25,
author={Otmane Azeroual},
title={Beyond Black Boxes: Adaptive XAI for Dynamic Data Pipelines},
booktitle={Proceedings of the 17th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management - Volume 2: KMIS},
year={2025},
pages={428-437},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013736100004000},
isbn={978-989-758-769-6},
}
in EndNote Style
TY - CONF
JO - Proceedings of the 17th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management - Volume 2: KMIS
TI - Beyond Black Boxes: Adaptive XAI for Dynamic Data Pipelines
SN - 978-989-758-769-6
AU - Azeroual O.
PY - 2025
SP - 428
EP - 437
DO - 10.5220/0013736100004000
PB - SciTePress