Deadline Date: 28 February 2026
1. Issue Introduction
Background:
The integration of Artificial Intelligence (AI) into engineering disciplines has revolutionized simulation methodologies, enabling faster and more accurate analyses. However, the adoption of complex AI models, particularly deep learning architectures, has introduced challenges related to their "black-box" nature. This opacity hinders the understanding of how models arrive at specific conclusions, raising concerns about transparency, interpretability, and trust, especially in safety-critical and regulatory-mandated applications.
Importance:
Explainable Artificial Intelligence (XAI) emerges as a pivotal paradigm to bridge the gap between AI's predictive capabilities and human understanding. By elucidating the decision-making processes of AI models, XAI enhances the interpretability of model outputs, allowing engineers, designers, and decision-makers to verify, trust, and act on AI-generated outcomes with confidence. Integrating XAI into numerical modeling workflows not only reinforces the transparency of engineering simulations but also ensures ethical alignment and fosters wider acceptance of AI in mission-critical engineering applications.
2. Aim and Scope
This Special Issue aims to explore the incorporation of XAI into numerical modeling pipelines to enhance the transparency, interpretability, and trustworthiness of engineering simulations. We invite original, high-quality contributions that:
Investigate novel XAI approaches tailored for computational engineering challenges.
Examine the integration of XAI with physics-based modeling techniques.
Address challenges and limitations in implementing XAI within engineering simulations.
Propose methodologies for validating and verifying AI-driven simulation outcomes.
Discuss the ethical implications and regulatory considerations of deploying XAI in engineering contexts.
3. Suggested Themes
Interpretable Deep Learning Models for Structural Failure Prediction in Finite Element Simulations
SHAP-Enhanced Surrogate Modeling for Real-Time Design Optimization in Aerospace Engineering
Improving Trust in Data-Driven Thermal Simulations Using LIME-Based Explainable Neural Networks
Physics-Informed XAI Models for Transparent Fluid Flow Simulation in Complex Geometries
Explainable Graph Neural Networks for Mechanical System Failure Diagnosis
Visualizing Attention Mechanisms in Transformer-Based AI for Multiphysics Simulation Pipelines
Model-Agnostic Interpretability of AI-Augmented Vibration Analysis in Civil Structures
Hybrid XAI Framework for CFD Simulations Using Symbolic Regression and Neural Operators
Transparent Control Decisions in Smart Manufacturing Using Reinforcement Learning and XAI
Saliency-Guided CNN Interpretability for Heat Distribution Modeling in Engineering Materials
Quantifying Uncertainty in AI-Based Load Prediction Systems with Explainable Bayesian Inference
Fuzzy Logic-Driven XAI for Safety-Critical Engineering Design Under Uncertainty
Explainable Decision Trees for Modeling and Diagnosing Fatigue in Composite Materials
Integration of Capsule Networks and XAI for Failure Mode Identification in Structural Component
1. Issue Introduction
The integration of Artificial Intelligence (AI) into engineering disciplines has revolutionized simulation methodologies, enabling faster and more accurate analyses. However, the adoption of complex AI models, particularly deep learning architectures, has introduced challenges related to their "black-box" nature. This ... show more
Editorial board
