This book examines how artificial intelligence can remain accurate, transparent, and dependable when deployed in real-world systems. Its central premise is that reliability must be engineered, measured, and maintained throughout the system lifecycle.
The book brings together federated learning, explainable AI, and physics-informed intelligence within a unified reliability framework. Federated learning enables organisations to develop shared models without exchanging sensitive data. Explainable AI makes model reasoning open to human examination. Physics-informed intelligence constrains learning through physical laws and operational knowledge. These approaches address different sources of failure and are therefore most effective when designed as complementary parts of the same system.
By connecting these methods with reliability engineering, the book shows how AI failures can be anticipated, assessed, contained, and reduced. Manufacturing, healthcare, and energy case studies illustrate how reliability metrics, human oversight, and technical evidence can be incorporated into model development and decision support. Readers will gain both the conceptual foundations and practical methods needed to build AI systems that people and organisations can responsibly depend on.
Produkteigenschaften
- Artikelnummer: 9783032484055
- Medium: Buch
- ISBN: 978-3-032-48405-5
- Verlag: Springer
- Erscheinungstermin: 20.05.2027
- Sprache(n): Englisch
- Auflage: Erscheinungsjahr 2027
- Serie: Springer Series in Reliability Engineering
- Produktform: Gebunden
- Format (B x H): 155 x 235 mm
- Ausgabetyp: Kein, Unbekannt
