Integration of digital twins and physical AI in cyber-physical systems

This article proposes an adjustable-error neural network (NN) approximator and incorporates it into the adaptive neural tracking controller design of uncertain nonlinear systems. Noted that the error between the unknown nonlinear function and the NN approximator cannot be adjusted under the traditional NN control framework, as it is solely determined by the selection of neurons, basis functions, and the estimation of the ideal weight vector. This inherent constraint compromises the precision of the NN approximation and the convergence accuracy of the tracking error. To improve the approximation accuracy of unknown nonlinear functions in adaptive neural control systems, an adjustable-error NN approximator is designed, in which the error between the approximator and the unknown nonlinear function can be adjusted by designed parameters. Based on the proposed NN approximator, an adaptive neural tracking controller is designed for a class of uncertain nonlinear systems, which achieves higher accuracy of the tracking error compared with traditional methods. The stability of the resulting closed-loop system is proved in the Lyapunov sense, and the convergence of the tracking error is also analyzed. The effectiveness of the proposed scheme is verified by simulation and experiment.

Triple-negative breast cancer (TNBC) is the most aggressive subtype of breast cancer, characterized by the absence of estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 expression. The absence of molecular targets in TNBC limits treatment options and contributes to increased rates of recurrence, metastasis, and resistance to conventional therapies. TNBC, however, is rich in tumor-infiltrating lymphocytes; hence, elevated programmed death-ligand 1 expression makes these tumors amenable to immunotherapy. A variety of antibodies and immunomodulatory drugs are being explored for TNBC treatment, this review specifically focuses on design of immunomodulatory peptides for TNBC treatment. This review comprehensively discusses the peptide-based approaches for immunomodulating tumor microenvironment (TME) of TNBC and to enhance antitumor immune response. The peptide-mediated modulation of innate immune cells including tumor-associated macrophages, neutrophils, dendritic cells, and natural killer cells, as well as T cells of the adaptive immune system is explored in detail. The applications of peptides as immune checkpoint inhibitors and highlights of emerging strategies that employ peptides to induce immunogenic cell death to stimulate antitumor immunity are discussed. Immunogenic cell death inducing peptides promote the release of immunogenic signals such as damage-associated molecular patterns from dying cancer cells, which further activate dendritic cells, T cells, and neutrophils, thereby reshaping the TME to support robust antitumor immunity. These strategies underscore the transformative potential of peptide therapeutics to harness the immune system and reshape TME, offering a promising avenue for more effective and durable TNBC treatment. SIGNIFICANCE STATEMENT: With growing interest in peptides as tumor microenvironment modulator, this review provides an in-depth analysis of interactions and crosstalk between immune and tumor cells and explores therapeutic potential of peptides in modulating immune cell signaling pathways, with ultimate impact as anticancer agents for triple-negative breast cancer.

Growing demands in the field of energy storage materials have driven the need for ways of manufacturing more adaptive, efficient, and sustainable systems. This has increased interest in advanced manufacturing technologies that enable both structural programmability, and functional responsiveness. 4D printing (4DP), a step in evolution of additive manufacturing, uses stimuli-responsive smart materials (such as shape memory polymers, hydrogels, nanocomposites, and metal oxides) to fabricate components that are capable of time-dependent dynamic reconfiguration. This study investigates the intersection of 4DP technology and energy storage systems by critically evaluating the materials, processes, and device-centric applications of 4DP in batteries, supercapacitors, and fuel cells. This study categorizes electrochemical storage types, their material requirements, and current synthesis methods, identifying key limitations in energy efficiency, waste, and adaptability. 4DP-compatible materials are thoroughly analyzed in terms of their printability, structural integrity, and functional performance under various stimuli. Multiple case studies demonstrate thermal actuation, shape recovery, and self-healing in energy devices. A comparative analysis was also conducted between 3D printing (3DP) technology and 4DP according to parameters such as energy consumption, material waste, flexibility, and scalability. Current technological barriers identified in the literature include low throughput, complexities in ink formulation, and postprint activation requirements, which are discussed along with emerging solutions. With this review, the authors position 4D printing as a potential alternative manufacturing strategy for energy storage systems, particularly in applications requiring programmable architecture and functional adaptability, while recognizing that substantial technical and scalability challenges remain. • 4D printing enables programmable, morphing energy storage device architectures. • Smart materials like SMPs, hydrogels, and MXenes enhance energy storage functions. • 4DP reduces post-processing waste but faces recycling and scalability challenges. • Comparative analysis shows 4DP improves adaptability over traditional techniques. • Multifunctional electrodes show enhanced performance via stimuli-responsive design.

Towards extreme application scenarios: perspectives on artificial intelligence-driven smart energy management systems

This paper addresses the challenge of estimating T1-T2 distributions from NMR measurements in well-logging applications, where the number of measurements is constrained by associated costs. Adaptive sensing is incorporated into the NMR inverse problem framework to improve estimation accuracy under experimental constraints, with a focus on two-dimensional T1-T2 experiments. This work proposes two adaptive measurement systems that dynamically select measurements to perform. For the case of T1-T2 experiments, it is the set of wait times (TW) that is chosen. The proposed adaptive measurements use the mutual information between the observations and the unknown relaxation time distribution to prioritize those measurements that are expected to yield the highest information gain. The proposed adaptive measurement system is evaluated by comparing its performance with a fixed measurement system that comprises a complete set of measured data, using the mean root mean square error (MRMSE) as the evaluation metric. It is demonstrated through experimental results that improved estimation accuracy is achieved with a reduced number of measurements, and hence measurement cost is reduced.

Adaptive type-3 fuzzy logic system for modeling wear loss and coefficient of friction in composite materials

As building regulations grow in complexity and digital design workflows become more integrated, the need for automated compliance checking (ACC) is intensifying. This review combines PRISMA with AI-assisted semantic retrieval and concept mapping, using Boolean and Deep Search to identify 88 peer-reviewed studies and broaden coverage across disciplines. The dual approach balances methodological rigour with the discovery capacity needed to surface studies missed by conventional keyword searches, enabling lifecycle-oriented synthesis. The paper synthesises recent advances across rule-based, ontology-driven, and AI-enhanced ACC systems, tracing a shift toward more flexible, lifecycle-aware compliance frameworks. It also introduces a conceptual map that visualise ACC processes across five lifecycle stages. Persistent barriers include interoperability gaps, limited post-construction support, and challenges in large-scale rule formalisation. Findings indicate the growing need for hybrid tools that support re-checking and traceability. The paper outlines future directions for transparent, adaptive, and jurisdiction-sensitive ACC systems in design automation and regulatory practice. • ACC systems are systematically reviewed using PRISMA and Deep Search. • A conceptual map illustrates an integrated ACC framework across the lifecycle. • Early rule-based ACC has evolved into AI, LLM, and ontology-driven systems. • Future directions focus on interoperability, re-checking, and rule scalability.

AFS-GNN: Adaptive and fast scheduling system for distributed GNN training

Light-driven host-guest supramolecular transport engineering an antibacterial trap switch.

Color-resolved light field shaping via diffractive-electronic U-shape network with wavelength-aware virtual branching.

Structural rigidities and resource constraints in South Africa's agri-business sector have created uneven innovation pathways, concentrating technological progress among dominant firms and widening disparities between large- and small-scale operations. These dynamics raise concerns about resilience and equitable growth in a sector critical to food security and economic development. The study aimed to investigate the heterogeneity of innovation and resilience pathways within South Africa's agri-business sector, focusing on how structural factors influence firms' innovative conduct and performance. Using secondary data from 532 firms collected through the Human Sciences Research Council's Agricultural Business Innovation Survey (2019–2021), the analysis employed Bayesian Structural Equation Modelling to examine structure–conduct–performance relationships, dimensionality reduction techniques to identify innovation clusters, and network analysis to assess connectivity and vulnerability within the sector. The findings revealed significant innovation heterogeneity, with clusters differentiated by firm size and strategic orientation. Network analysis uncovered a hub-and-spoke structure dominated by large firms, which, while efficient, exhibited low modularity and transitivity, making it vulnerable to systemic shocks. Bayesian SEM indicated a negative relationship between structural rigidity and innovative conduct, and a positive link between conduct and performance, suggesting that structural constraints hinder innovation while proactive conduct enhances turnover. Despite stable employment and rising turnover, innovation adoption remained concentrated in process improvements and soil fertility priorities, constrained by barriers such as water scarcity, limited finance, and restrictive policies. The sector thus reflects a duality between entrenched conventional practices and emerging niche innovations shaped by structural rigidities. These insights underscore the need for targeted, system-level interventions to address structural rigidities, strengthen innovation capabilities, and enhance resilience in South Africa's agri-business sector. Aligning with the Agriculture and Agro-processing Master Plan, policy efforts should focus on expanding and maintaining water infrastructure, providing financial incentives for digital and climate smart technologies including artificial intelligence applications, and fostering collaborative platforms to promote inclusive innovation and resilience across firm sizes. Such measures would help reduce network fragility, broaden participation in higher-value innovation activities, and support a more inclusive, competitive, and adaptive agri-food system. • Turnover grew between 2019 and 2021, employment stable, but innovation barriers persist. • Clusters by agribusiness size show heterogeneity. • Large agribusinesses dominate hub-and-spoke networks. • Structure hinders conduct and conduct boosts performance. • Policy needs water and AI incentives.

Investigating the spatial heterogeneity in distance decay of rail passenger transportation: empirical evidence in Guangdong Province, China

Adaptive inference through Bayesian and inverse Bayesian inference with symmetry bias in nonstationary environments.

Shell structures are renowned for their unparalleled stiffness-to-weight performance through their elegant curvature and material efficiency. This review provides a comprehensive synthesis of the foundational principles and emerging technologies driving the next generation of shell design. It begins by examining key geometric concepts such as minimal surfaces, funicular geometry and inverse design, which enable structurally efficient shapes with minimal material use. The discussion then moves to computational optimization strategies (including size, shape, and topology optimization) as well as data-driven and AI-assisted frameworks that accelerate design exploration and performance prediction. Reinforcement approaches are surveyed next, spanning conventional stiffeners and sandwich shells. These strategies are contextualized through applications in civil infrastructure, aerospace components, and biomedical implants. The review concludes by identifying key challenges such as imperfection sensitivity, fabrication constraints, and multi-objective trade-offs, while highlighting promising directions in hybrid multiscale modeling and sustainable shell design. Together, the topics covered provide an integrated roadmap for the development of intelligent, adaptive, and high-performance shell systems.

Enhanced optimization of the diagnosis of liver disease with an intelligent adaptive neuro-fuzzy inference system

As urban transportation evolves, shared e-mobility is increasingly recognized as a socio-technical system shaping urban equity, social inclusion, and mobility behaviour. However, existing platforms often lack multimodal integration and user-centric adaptability, limiting their ability to address diverse and behaviourally heterogeneous travel preferences. This study proposes a cloud-based shared e-mobility platform integrating docking electric cars, e-bikes, and e-scooters with large language model assistance for natural-language preference interpretation. The system enables human-centred decision-making through lexicographic multi-objective route optimization. The platform is evaluated using 500 expert-annotated queries, assessing both preference alignment and optimization performance. Results show that gpt-4.1 achieves the highest semantic alignment score (0.928) and best route quality measured by a Mean Optimality Gap of 18.62%. To jointly capture alignment and optimization performance, we introduce the Combined Alignment–Performance Metric, under which gpt-4o achieves the highest score (1.0594). Optimization experiments demonstrate high system robustness under varying traffic conditions, as key metrics such as travel time ( p = 0 . 771 ), risk ( p = 0 . 341 ), and walking distance ( p = 0 . 153 ) show no statistically significant differences. Furthermore, the platform exhibits significant scalability, where increasing e-hub density from 20 to 100 stations reduced the mean travel time from 1771 ± 497 s to 955 ± 343 s ( p < 0 . 001 ). These findings contribute to interdisciplinary transportation research by linking optimization with human-centred mobility analysis, offering actionable insights for equity-aware urban planning, inclusive mobility system design, and policy development supporting adaptive and sustainable transport systems. • Cloud platform enables multimodal routing across e-bikes, e-scooters and e-cars. • LLMs translate user travel preferences into lexicographic optimization priorities. • Eight LLMs benchmarked for preference alignment and routing performance. • Compact models achieve comparable route quality with faster response time. • Denser e-hub networks significantly reduce average multimodal travel time.

Bridges are critical components of transportation infrastructure, and ensuring their structural integrity is essential for public safety and efficient mobility. Traditional inspection methods largely rely on manual evaluation, which is time-consuming, labor-intensive, and susceptible to human error, partic-ularly when detecting small or early-stage cracks. To address these limitations, this paper proposes an adaptive vision-based system for the early detection of micro-cracks in bridges using advanced deep learning and computer vision techniques. The proposed framework integrates Convolutional Neural Networks (CNNs) for feature extraction, YOLO for real-time object detection, and U-Net for precise image segmentation, enabling accurate identification and localization of cracks from images captured via drones or fixed cameras. The system is designed to operate effectively under diverse environmental conditions, including low lighting, shadows, and noise, ensuring robustness in real-world scenarios. Additionally, the use of lightweight models allows deployment on edge devices, facilitating real-time processing and reducing computational overhead. Experimental results demonstrate that the proposed system achieves high detection accuracy and effectively highlights crack regions for detailed analysis. By automating the inspection process, the approach minimizes human intervention, enhances operational safety, and supports timely maintenance decisions. Overall, the proposed system provides a practical, scalable, and efficient solution for intelligent bridge monitoring and long-term infrastructure management

With the rapid growth of intelligent systems and personalized user experiences, emotion-aware applications have gained significant attention. This project presents a Facial Emotion-Based Song Recommendation System that automatically detects a user's emotional state through facial expressions and recommends music accordingly. The system utilizes computer vision techniques and deep learning models to analyze real-time facial inputs captured via a webcam. The proposed model employs algorithms based on Computer Vision and Deep Learning, particularly Convolutional Neural Networks (CNNs), to classify emotions such as happiness, sadness, anger, surprise, fear, and neutrality. Once the emotion is identified, the system maps it to a curated music database and suggests songs that align with the detected mood, enhancing user engagement and emotional well-being. The system integrates facial detection frameworks like OpenCV and machine learning libraries such as TensorFlow or Keres for model training and deployment. The recommendation engine may use content-based filtering or emotion-tagged playlists to provide relevant suggestions. This approach demonstrates how affective computing can be leveraged to create adaptive and intuitive music recommendation systems. The implementation aims to improve user satisfaction by delivering a seamless and personalized music experience based on real-time emotional analysis

Online examinations require reliable mechanisms to ensure academic integrity; however, many existing systems rely on rigid rule-based approaches that often generate false alerts and fail to account for individual behavioral differences. This paper presents ScoreHunt, an AI-based adaptive cheating detection system that analyzes personalized student behavior using computer vision and keystroke dynamics. The proposed system establishes a behavioral baseline for each student and continuously compares it with real-time activities, including facial presence, gaze patterns, and typing behavior. A multi-indicator validation mechanism is employed to improve detection accuracy while minimizing false positives. Additionally, the system provides automated alerts, detailed activity logs, and an administrative dashboard for efficient monitoring and analysis. Experimental evaluation conducted across 15 test cases demonstrates stable performance with reduced false positive rates, indicating the effectiveness of the proposed approach in enhancing the reliability and fairness of online examination systems.