Explainable AI for Evolutionary Computation

Erscheinungsjahr 2025
ISBN: 978-981-962539-0
Verlag: Springer Nature Singapore

Buch, Englisch, 195 Seiten, Format (B × H): 160 mm x 241 mm, Gewicht: 512 g

Reihe: Natural Computing Series

This book explores the intersection between explainable artificial intelligence (XAI) and evolutionary computation (EC). In recent years, the fields of XAI and EC have emerged as vital areas of study within the broader domain of artificial intelligence and computational intelligence. XAI seeks to address the pressing demand for transparency and interpretability in AI systems, enabling their decision-making processes to be scrutinised and trusted. Meanwhile, EC offers robust solutions to complex optimisation problems across diverse and challenging domains, drawing upon the principles of natural evolution. While each field has made significant contributions independently, their intersection remains an underexplored area rich with transformative potential.

This book charts a path towards advancing computational systems that are transparent, reliable, and ethically sound. It aims to bridge the gap between XAI and EC by presenting a comprehensive exploration of methodologies, applications and case studies that highlight the synergies between these fields. This book will serve as both a resource and an inspiration, encouraging researchers and practitioners within XAI and EC, as well as those from adjacent disciplines, to collaborate and drive the development of intelligent computational systems that are not only powerful but also inherently trustworthy.

Zielgruppe

Research

Autoren/Hrsg.

Kononova, Anna V.

van Stein, Niki

Fachgebiete

• Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz

Inhaltsverzeichnis

Introduction.- Overview of XAI methods and their applicability for EC.- Feature Importance and Sensitivity Analysis.- Explainable Landscape Analysis.- EC based XAI methods.- XAI for Benchmarking Black Box Metaheuristics.- XAI for Automatic Algorithm Configuration.- XAI for Multi Criteria Decision Making.-Applications of XAI in EC.

Über Autor(innen)

Niki van Stein: Dr. Niki van Stein is an Assistant Professor of Explainable Artificial Intelligence (XAI) and Evolutionary Computing at Leiden University, The Netherlands. She leads the XAI research group, part of the Natural Computing Cluster within the Leiden Institute of Advanced Computer Science (LIACS). Her work focuses on integrating XAI principles into complex computational systems, particularly within evolutionary computation and optimization frameworks.
With a background in predictive maintenance, time-series analysis, machine learning and optimization, Dr. van Stein brings a multidisciplinary approach to her research. She has authored over 90 peer-reviewed publications on algorithmic interpretability and evolutionary computation methods and is actively involved in the academic community, serving as program chair of the IJCCI and EXPLAINS conferences, as editorial board member of the Evolutionary Computation journal and numerous other contributions. Dr. van Stein is passionate about making AI systems more transparent and accessible while pushing the boundaries of natural computing to solve real-world problems.

Anna V. Kononova: Dr Anna V. Kononova is an Assistant Professor of Efficient Heuristic Optimisation (EcHO) at Leiden University, the Netherlands. She leads the EcHO research group, part of the Natural Computing Cluster within the Leiden Institute of Advanced Computer Science. Her research focuses on achieving order-of-magnitude efficiency improvements in solving heuristic optimisation problems by integrating elements of machine learning and robust algorithmic design.
With expertise in heuristic optimisation, machine learning, algorithm analysis and applied problem-solving, Dr Kononova takes a multidisciplinary approach to address complex challenges at an appropriate level of abstraction. She has authored over 75 peer-reviewed publications, contributing significant insights into the behaviour and performance of optimisation algorithms across a variety of contexts. Dr Kononova actively contributes to the scientific community, serving as the editorial board member of the Evolutionary Computation journal and organising leading conferences in the field, such as PPSN, EMO, GECCO and FOGA.
Dedicated to bridging the gap between theoretical research and practical implementation, Dr Kononova strives to make heuristic optimisation methods more accessible and impactful. Her work continues to advance the field, driving innovation and progress in natural computing.