Preface for “Artificial Intelligence and Algorithms with Their Applications”

Artificial Intelligence (AI) has become a driving force of advancements in science, engineering, and society. While recent progress has been fueled by large-scale data and massive computing resources, the long-term success of AI systems fundamentally depends on sound mathematical modeling, robust algorithms, and principled theoretical foundations. This Special Issue, Artificial Intelligence and Algorithms with Their Applications, brings together eleven peer-reviewed papers that collectively highlight the role of mathematical rigor and algorithmic innovation in modern AI research.

1. Perception and Computer Vision

An overarching theme of this collection is deep learning–based perception and computer vision. In Contribution 1, Turkmen and Akgun propose a hybrid UNet incorporating Transformer attention and a novel perceptual loss function for monocular depth estimation, achieving state-of-the-art accuracy on benchmark datasets. Complementing this, Hu (Contribution 2) provides a comprehensive mathematical survey of deep edge-detection algorithms, tracing the progression from early convolutional models to recent attention-based approaches and generative paradigms. This survey offers a valuable reference for researchers in symbolic and low-level vision.

Interpretability in vision models is addressed by Chavarro et al. (Contribution 3), who conduct a comparative evaluation of Class Activation Mapping (CAM) techniques. This work emphasizes the importance of eXplainable AI (XAI) in critical application domains, such as plant disease detection. Additionally, Yang et al. (Contribution 11) exploit local depth properties to enhance 3D feature matching.

2. Decision-Making and Control

Another salient focus of the Special Issue is learning-based decision-making and control. Autonomous decision-making under uncertainty has a wide application domain. Mei et al. (Contribution 4) have developed a deep reinforcement learning framework for autonomous air-combat maneuver generation, combining hierarchical policies, recurrent modeling, and self-play training to address challenges in complex, dynamic, and partially observable environments. From an industrial perspective, Contribution 5 introduces an AI-driven predictive maintenance framework supported by digital twin technology. This paper illustrates how incorporating machine learning, optimization, and system modeling can enhance reliability and efficiency in smart manufacturing.

3. Security and Trustworthy AI

Secure and trustworthy AI systems form a central pillar of modern research. Abdel-Basset et al. (Contribution 6) present a blockchain-integrated federated learning approach for pandemic diagnosis in smart cities, addressing data privacy and decentralization. Relatedly, Reddy et al. (Contribution 7) develop a hybrid fuzzy trust mechanism for IoT-embedded systems, while Kiran et al. (Contribution 10) enhance IoT security via multi-level blockchain and adaptive clustering.

4. Pattern Matching and Semantic Analysis

Finally, the Special Issue broadens its scope through pattern matching, text mining, and semantic analysis. Na et al. (Contribution 8) present an efficient approximate order-preserving pattern matching algorithm, which extends classical string-matching techniques to manage noisy and real-valued time-series data. In Contribution 9, a mathematically grounded framework is proposed for extracting spatial information from non-spatial text using conceptual spaces, entropy measures, and spatial statistics, highlighting the applicability of AI algorithms to natural language processing and semantic inference.

5. Conclusions

Collectively, the eleven papers in this Special Issue showcase the breadth, depth, and mathematical foundations of contemporary AI research. By addressing perception, learning, optimization, interpretability, and text analysis, this collection underscores the essential role of algorithms and applied mathematics in advancing reliable, explainable, and high-performance artificial intelligence systems.