BackgroundThe lack of osteoporosis treatment initiation following fragility fractures is a recognized gap, particularly in primary care. Primary care physicians’ (PCPs) barriers to treatment, such as uncertainties in investigation, initiation, and concerns about drug side effects, remain challenging. It is also unclear whether knowledge gaps and barriers vary by region or if active learning platforms are more effective than passive methods in improving treatment rates, and how PCP demographics influence learning outcomes. With time constraints, PCPs are increasingly using online platforms for continuing professional development, and the interactive online Community Fracture Capture (CFC) tool has emerged as a promising alternative to traditional methods. Our CFC pilot study tested this program’s design and content, revealing its potential effectiveness.ObjectiveThe study aimed to assess the operational characteristics, educational effectiveness, and acceptability of the interactive online CFC model in enhancing Australian PCPs’ knowledge and skills in community-based fracture treatment. Additionally, it sought to examine how PCPs’ knowledge and treatment gaps relate to their demographic characteristics and clinical practice backgrounds.MethodsThe CFC Learning Hub is a secure, adaptable online platform that promotes community learning. It includes an interactive forum where participants share case studies and engage in discussions with bone specialists and senior PCP facilitators. The hub also offers a knowledge repository and allows participants to post inquiries. Online surveys and back-end analytics track baseline knowledge, activity levels, and improvements in knowledge and confidence over time, offering insights into participants’ learning and program development.ResultsFour 6-week small-group cycles involved 55 PCPs, with over 80% working in metropolitan-based practices and a median (IQR) of 22 (16-34) years in practice. Topic modules covered osteoporosis diagnostics, treatment, monitoring, and challenging conditions, using a multidisciplinary approach with participant case studies. A total of 35 (64%) PCPs provided evaluation data, with 86% (n=30) joining to learn from experts or improve patient management and 83% (n=29) being satisfied with the content. Preferred learning methods included small group learning (n=13, 37%), live webinars (n=9, 26%), interactive learning (n=7, 20%), and on-demand videos (n=6, 17%), and 57% (n=20) found the platform easy to use. The most popular access times were evening (n= 23, 66%) and weekends (n=10, 29%). At completion, 89% (n=31) would recommend the training, and 78% (n=22 out of 28 respondents to the postprogram expectations meeting survey) were fully satisfied that their training needs were met, with 22% (n=6) partly satisfied. In addition, following the course completion, almost everyone reported being confident or very confident in managing osteoporosis.ConclusionsThe CFC program was created by bone specialists, PCPs, software engineers, and information technology specialists. This collaboration produced a user-friendly, case-based, interactive, time-flexible, and highly acceptable program bridging investigation and management gaps in osteoporosis. It is customized to address challenges faced by PCPs and is potentially relevant for implementation in a wide range of fields, both health-related and others.

Successful delivery of lung cancer radiotherapy is hindered by respiratory motion, low soft-tissue contrast and anatomical variabilities, often compromising precision. Magnetic Resonance Image-guided Radiotherapy (MRgRT) has emerged as a promising approach, particularly with hybrid MR-Linac systems that offer superior soft-tissue visualization and enable online adaptive radiotherapy (online MRgART). This review synthesizes current evidence for online MRgART in lung cancer and examines emerging roles for quantitative MRI (qMRI). A PubMed search covering the period from January 2020 to September 2025 identified 19 studies, 3 of which focused specifically on quantitative imaging. Online MRgART consistently demonstrated workflow feasibility, frequent online adaptation, improved target coverage while respecting Organs-At-Risk (OARs) constraints and encouraging Local Control (LC) with low high-grade toxicity. qMRI on MR-Linacs, most commonly Diffusion-Weighted Imaging (DWI) and cine-MRI-derived ventilation/perfusion mapping, showed feasibility and early signals for treatment adaptation, toxicity prediction and response assessment. qMRI studies integrated in online MRgART for lungs are, at present, extremely limited; nevertheless, establishing a clear snapshot of the current state-of-the-art is essential, as this topic is expected to become highly prevalent and of particular interest in the near future. To our knowledge, this is the first review centered on online MRgART for lung tumors, with a dedicated subsection summarizing the nascent evidence on qMRI. Looking ahead, integrating AI-driven motion compensation, auto-segmentation and adaptive replanning with qMRI-enabled biomarkers could standardize workflows and accelerate truly personalized online MRgART. Prospective multi-center studies are needed to validate biomarkers and demonstrate clinical benefit.

The nursing profession was established in a spirit of individual policy entrepreneurship, yet this identity has been sparsely activated despite a recognized social contract with the public as a trusted advocate and a history of nurse policy entrepreneurs. The Nurse Policy Entrepreneur Café (NPEC) is an unaffiliated volunteer group that focuses on skill development, networking, innovative approaches to identifying barriers, and offering support and resources to empower interested nurses to create their personal and professional development strategy to become effective partners in health-care policy development, reform, and implementation. The NPEC was established in 2022, utilizing an online adaptation of The World Café Method to paint a picture of the Nurse Policy Entrepreneur identity, and has since expanded to include resources such as presentations, blogs, a newsletter, and support for global participant networking, supporting the needs of all nurses and nursing students worldwide.

In struggling against invasive species ravaging riverscape ecosystems, gaps in dispersal pathway knowledge and fragmented approaches across scales have long stalled effective riparian management worldwide. To reduce these limitations and enhance invasion management strategies, selecting appropriate alien species as models for in-depth pathway analysis is essential. Alternanthera philoxeroides (Mart.) Griseb. (alligator weed) emerges as an exemplary model species, boasting an invasion record of around 120 years spanning five continents worldwide, supported by genetic evidence of repeated introductions. In addition, the clonal reproduction of A. philoxeroides supports swift establishment, while its amphibious versatility allows occupation of varied riparian environments, with spread driven by natural water-mediated dispersal (hydrochory) and human-related vectors at multiple scales. Thus, leveraging A. philoxeroides, this review proposes a comprehensive multi-scale framework, which integrates monitoring with remote sensing, environmental DNA, Internet of Things, and crowdsourcing for real-time detection. Also, the framework can further integrate, e.g., MaxEnt (Maximum Entropy Model) for climatic suitability and mechanistic simulations of hydrodynamics and human-mediated dispersal to forecast invasion risks. Furthermore, decision-support systems developed from the framework can optimize controls like herbicides and biocontrol, managing uncertainties adaptively. At the global scale, the dispersal paradigm can employ AI-driven knowledge graphs for genetic attribution, multilayer networks, and causal inference to trace pathways and identify disruptions. Based on the premise that our multi-scale framework can bridge invasion ecology with riverscape management using A. philoxeroides as a model, we contend that the implementation of the proposed framework tackles core challenges, such as sampling biases, shifting environmental dynamics, eco–evolutionary interactions using stratified sampling, and adaptive online algorithms. This methodology is purposed to offer scalable tools for other aquatic invasives, evolving management from reactive measures to proactive, network-based approaches that effectively interrupt dispersal routes.

Advanced cognition can be measured from the human brain using brain-computer interfaces (BCIs). Integrating these interfaces with computer vision techniques, which possess efficient feature extraction capabilities, can achieve more robust and accurate detection of dim targets in aerial images. However, existing target detection methods primarily concentrate on homogeneous data, lacking efficient and versatile processing capabilities for heterogeneous multimodal data. In this article, we first build a brain-eye-computer-based object detection system for aerial images under few-shot conditions. This system detects suspicious targets using region proposal networks (RPNs), evokes the event-related potential (ERP) signal in electroencephalogram (EEG) through the eye-tracking-based slow serial visual presentation (ESSVP) paradigm, and constructs the EEG-image data pairs with eye movement data. Then, an adaptive modality balanced online knowledge distillation (AMBOKD) method is proposed to recognize dim objects with the EEG-image data. AMBOKD fuses EEG and image features using a multihead attention module, establishing a new modality with comprehensive features. To enhance the performance and robust capability of the fusion modality, simultaneous training and mutual learning between modalities are enabled by end-to-end online KD (OKD). During the learning process, an adaptive modality balancing module is proposed to ensure multimodal equilibrium by dynamically adjusting the weights of the importance and the training gradients across various modalities. The effectiveness and superiority of our method are demonstrated by comparing it with existing state-of-the-art methods. Additionally, experiments conducted on public datasets and real-world scenarios demonstrate the reliability and practicality of the proposed system and the designed method. The dataset and the source code can be found at: https://github.com/lizixing23/AMBOKD.

We present H-NGPCA, a hierarchical clustering algorithm for data streams that integrates an adaptive unit number growth and local dimensionality control. Unlike existing algorithm, H-NGPCA combines the characteristics of centroid-based, model-based and hierarchical clustering. H-NGPCA builds a hierarchical structure of local Principal Component Analysis (PCA) units, where each unit is a hyper-ellipsoid whose shape is updated by a neural network-based online PCA method. The re-positioning of each unit is handled by Neural Gas, a centroid-based clustering algorithm. In the hierarchical tree structure, new units are created in a branch if suggested by a splitting criterion. In addition, each unit determines its own dimensionality based on the data represented by the unit. In extensive benchmarks, H-NGPCA not only surpasses all competing online algorithms with adaptive unit numbers but also achieves competitive performance with state-of-the-art offline methods, reaching an average NMI = 0.87 and CI = 0.26. This demonstrates that H-NGPCA achieves both online adaptability and offline-level accuracy.

Abstract A control strategy is proposed for a trapezoidal back-EMF permanent magnet synchronous motor (commonly referred to as a brushless DC motor) driven by a four-switch, three-leg inverter. The motor is regulated using field-oriented control (FOC) with an outer speed loop governed by a Dung-Beetle Optimization (DBO) algorithm–based Artificial Neural Network tuned Proportional–Integral–Derivative (D-BAP) controller. The PID gains are optimized at discrete reference speeds ranging from 50 rpm to 3000 rpm under three load conditions (low, medium, and high) using the Integral of Time-weighted Absolute Error (ITAE) criterion. The optimized gain sets, along with the corresponding reference speed, actual speed, battery current, and battery voltage, are compiled into an extensive dataset used to train the ANN for adaptive online tuning. The trained network continuously updates the controller gains in real time, enabling robust adaptation to variations in mechanical load and DC-link voltage. The proposed D-BAP controller is benchmarked against conventional PI, fuzzy–PI, Model Predictive Control (MPC), and Sliding Mode Control (SMC), with both simulation and experimental results demonstrating superior speed tracking, minimal steady-state error, and enhanced robustness under nonlinear operating conditions.

Recently, wireless sensor networks (WSNs) have been widely integrated in critical applications such as environmental monitoring, smart cities, and modern healthcare for remote patient monitoring and data collection. This makes WSNs increasingly susceptible to security threats, including eavesdropping, jamming, sybil, data injection, routing, senor node capture, malicious intrusion attacks etc., therefore maintaining integrity, confidentiality, and availability of sensitive data and preserving privacy become a challenge. Existing mechanisms do not integrate threat detection, privacy preservation, and adaptability to evolving threats leading to security breaches in the left-out security requirements. This paper proposes an ensemble-based threat detection mechanism (FAL-ELeM-IDS) with privacy-awareness and adaptability to evolving threats for WSNs-based healthcare systems. The ensemble consists of Online Random Forest, Online AdaBoost, Support Vector Machine, Neural Network, and XGBoost to ensure detection high accuracy and low false-positives. Federated Learning combined with ensemble technique to provide confidentiality and a combined Online Adaptive Boosting and Online Random Forests algorithms to provide adaptability. The proposed model trained on a real-world healthcare sensor dataset demonstrates its superiority in performance compared to conventional models. An accuracy of 97.8%, a recall of 97%, precision of 98%, and F1-score of 97.5%, was achieved outperforming individual models by significant margins, showing that the model is accurate and reliable in detecting threats. This mechanism implies enhanced system security and privacy, timely threat mitigation ensuring patient safety, and boost in public acceptance for sensor-based healthcare services. Overall, this work contributes a scalable, privacy-aware, and adaptive threat detection mechanism suitable for integration in the sensitive healthcare applications.

Non-Player Characters (NPCs) play a critical role in shaping player experience in modern digital games, yet traditional NPC behaviour is largely driven by static, rule-based logic that lacks adaptability and realism. Such approaches often result in predictable and exploitable gameplay, limiting long-term player engagement. This project presents Neo NPC, a reinforcement learning–based framework for developing adaptive NPC behaviour in a 2D fighting game environment. The proposed system leverages offline reinforcement learning to train intelligent NPC agents that exhibit progressively sophisticated combat strategies across predefined difficulty levels, namely Easy, Medium, and Hard. The study focuses on designing a structured training and deployment pipeline that decouples model learning from real-time gameplay execution. Gameplay environments are modelled to generate state-action-reward trajectories, which are used to train NPC policies using the Proximal Policy Optimization (PPO) algorithm. Trained policies are periodically evaluated, versioned, and categorized into difficulty tiers based on quantitative performance metrics such as win rate, damage efficiency, and survival time. These validated models are then integrated into the game engine through a modular inference layer, enabling real-time decision-making without modifying core game logic. Experimental results demonstrate clear behavioural differentiation across difficulty levels, with higher-tier models exhibiting improved defensive responses, reduced vulnerability to repetitive player strategies, and increased action diversity. Human playtesting further confirms that the adaptive NPCs provide a more challenging and engaging gameplay experience compared to traditional scripted opponents. The proposed approach highlights the effectiveness of reinforcement learning in producing scalable, reusable, and intelligent NPC behaviour. Future extensions of this work include online adaptation, multi-agent selfplay, and transfer of trained models to more complex game environments.

As an actuation mechanism for achieving precision attitude control in aircraft, the electromechanical actuator (EMA) plays a critical role in ensuring flight safety and stability. However, the EMA is subject to unmeasurable unknown disturbances that act through mismatched channels relative to the system’s control input. To address this, this paper employs feedback linearization to transform the existing model. The transformed model effectively recasts the unknown disturbance into the same channel as the control input, thereby enabling active disturbance rejection via control law design. Furthermore, to overcome the challenge of immeasurable disturbances, an extended state observer (ESO) is designed to estimate the unknown disturbance; the estimated value is then directly utilized in the control law synthesis. Subsequently, a fuzzy logic system (FLS) is developed to perform real-time online adaptation and optimization of the controller parameters. Finally, extensive simulation results are provided to validate the effectiveness of the proposed algorithm.

This study investigates the association between social media use and adolescent behavior among students at Muhammadiyah 2 Sidoarjo High School, East Java, Indonesia, addressing the limited empirical evidence from faith-based secondary education within urban, peripheral contexts. The research seeks to identify patterns of adolescents’ social media engagement and to examine how variations in usage intensity correspond to behavioral tendencies. A quantitative correlational design was applied, utilizing primary data obtained through an online questionnaire administered to 123 active students from grades X and XI. Descriptive statistics and simple linear regression were employed to evaluate the strength and direction of the relationship between variables. The findings demonstrate a positive and statistically significant directional association (B = 0.255, p < 0.05) between social media use and adolescent behavior, indicating behavioral amplification rather than normative improvement, whereby increased engagement corresponds to more pronounced expressions of both constructive and maladaptive behaviors. Nevertheless, the model’s explanatory power remains modest (Adjusted R² = 12.4%), suggesting that adolescent behavior is predominantly shaped by additional contextual factors, including family environment, religiosity, and peer dynamics. Overall, the results underscore the dual role of social media as a medium for educational engagement and a potential source of behavioral risk, reinforcing the need for structured digital literacy initiatives and guided supervision to foster adaptive online practices among adolescents.

Deep learning has become a popular topic among scholars and has attracted widespread attention. However, deep learning methods typically require large datasets to determine model parameters and can only process data in batches. To address the challenges of deep learning models, which rely on batch data and struggle to adapt to industrial streaming data scenarios in gearbox fault diagnosis, this study proposes an online gearbox fault diagnosis method based on a DeepWalk graph embedding-enhanced extreme learning machine (ELM) approach. The method constructs a graph structure in real time for each newly collected vibration signal, uses DeepWalk for unsupervised embedding learning, and extracts low-dimensional features with strong discriminative power. These features are then input into the ELM classifier to achieve adaptive fault type recognition and online incremental model updates. This method does not require historical data to be retrained, thus effectively overcoming the bottleneck of batch retraining and significantly improving diagnostic efficiency and resource utilization. The experimental results show that, under various operating conditions, the proposed method achieves fast and accurate diagnosis of multiple gearbox fault types, with an average accuracy consistently above 95%, thereby demonstrating excellent engineering applicability and real-time performance.

Fluorescence molecular tomography (FMT) is a promising medical imaging technology with the ability to quantitatively reconstruct the three-dimensional distribution of fluorescently labeled probes in vivo. However, due to the strong scattering properties of biological tissues, conventional reconstruction methods encounter challenges such as low reconstruction accuracy and high computational complexity. Here, an adaptive online variational Bayesian method based on the normal-generalized inverse Gaussian (NGIG) prior is proposed. This method reduces computational complexity while ensuring that the globally optimal solution is maintained. Specifically, by utilizing variational inference, the optimization of the objective function is converted into a convex optimization problem that minimizes the variational lower bound, effectively reducing the function's complexity. Furthermore, to accurately capture the prior distribution, the NGIG prior is introduced. It imposes probabilistic constraints on the sparsity structure. This approach alleviates the adverse effects caused by overly strict sparsity constraints. In addition, the adaptive gradient algorithm (Adagrad) is employed to dynamically adjust the parameter learning rate, thereby preventing the algorithm from becoming trapped in local optima during the posterior inference process. The effectiveness of the proposed method is validated through numerical simulations and fluorescence source implantation experiments. The results show that the adaptive online variational Bayesian (AOVB)-NGIG method achieves superior performance in both fluorescence source localization and shape recovery. The minimum localization error is 0.243 mm, accompanied by a dice coefficient of 0.889. Meanwhile, the root mean square error and relative intensity error remain relatively low, indicating that the reconstructed results are the closest to the actual light source. These outcomes demonstrate that AOVB-NGIG can reliably reconstruct the spatial characteristics of the fluorescence source with high accuracy. This study is expected to advance the preclinical and clinical applications of FMT in early tumor detection.

This paper develops the first online algorithms for estimating the spectral density function — a fundamental object of interest in time series analysis — that satisfies the three core requirements of streaming inference: fixed memory, fixed computational complexity, and temporal adaptivity. Our method builds on the concept of forgetting factors, allowing the estimator to adapt to gradual or abrupt changes in the data-generating process without prior knowledge of its dynamics. We introduce a novel online forgetting-factor periodogram and show that, under stationarity, it asymptotically recovers the properties of its offline counterpart. Leveraging this, we construct an online Whittle estimator, and further develop an adaptive online spectral estimator that dynamically tunes its forgetting factor using the Whittle likelihood as a loss. Through extensive simulation studies and an application to ocean drifter velocity data, we demonstrate the method’s ability to track time-varying spectral properties in real-time with strong empirical performance.

The sustainability of blended learning in private universities has become a critical concern following the Covid-19 pandemic, as this learning model combines face-to-face and online methods, offering flexibility and adaptability to technological developments. Case studies at Esa Unggul University (UEU), Mercu Buana University (UMB), and Tarumanagara University (UNTAR) provide a concrete overview of the scope of blended learning implementation and the challenges faced in maintaining its sustainability in private universities. This research aims to examine how the implementation of blended learning in private universities, especially at Esa Unggul University (UEU), Mercu Buana University (UMB), and Tarumanagara University (UNTAR). This research uses a case study approach with mixed methods to examine sustainability. Blended learning at three private universities in Jakarta: Esa Unggul University, Mercu Buana University, and Tarumanagara University. Data were collected through interviews with academic and administrative stakeholders and questionnaires with students in employee classes, supported by secondary data such as online learning documentation. Analysis was conducted thematically and quantitatively using SPSS, resulting in a comprehensive overview of the supporting factors and challenges in the implementation blended learning at the three universities. The research results show that implementing blended learning post-pandemic requires system support, digital skills, and institutional commitment to address challenges and maintain sustainability. Adaptation is carried out through development strategies, strengthening human resources, and optimizing learning methods. Sustainability prospects depend on technological readiness, teaching quality, and student engagement, although several aspects such as technical skills and feedback, need improvement.

BackgroundElectronic medical records (EMRs) are a potentially rich source of information on an individual’s health care providers’ clinical activities. These data provide an opportunity to tailor web-based learning for health care providers to align closely with their practice. There is increasing interest in the use of EMR data to understand performance and support continuous and targeted education for health care providers.ObjectiveThis study aims to understand the feasibility and acceptability of harnessing EMR data to adaptively deliver a web-based learning program to early-career physicians.MethodsThe intervention consisted of a microlearning program where content was adaptively delivered using an algorithm input with EMR data. The microlearning program content consisted of a library of questions covering topics related to best practice management of common emergency department presentations. Study participants were early-career physicians undergoing training in emergency care. The study design involved 3 design cycles, which iteratively changed aspects of the adaptive algorithm based on an end-of-cycle evaluation to optimize the intervention. At the end of each cycle, an online survey and analysis of learning platform metrics were used to evaluate the feasibility and acceptability of the program. Within each cycle, participants were recruited and enrolled in the adaptive program for 6 weeks, with new cohorts of participants in each cycle.ResultsAcross each cycle, all 75 participants triggered at least 1 question from their EMR data, with the majority triggering 1 question per week. The majority of participants in the study indicated that the online program was engaging and the content felt aligned with clinical practice.ConclusionsThe use of EMR data to deliver an adaptive online learning program for emergency trainees is both feasible and acceptable. However, further research is required on the optimal design of such adaptive solutions to ensure training is closely aligned with clinical practice.

System-On-Chip Systems platforms generate immense volumes of runtime activity logs that hold rich information about system behaviour, reliability, and fault conditions. Traditional log analysis techniques, however, are unable to efficiently unveil the underlying behavioural patterns and temporal dependencies necessary for effective anomaly detection. This work proposes a machine learning-driven framework for unveiling abstract behavioural patterns from SoC runtime activity logs with the synergistic combination of temporal graph embeddings and clustering-based abstraction and anomaly detection procedures. The scheme was tested on a RISC-V-based SoC prototype, using runtime traces under nominal and fault-injected scenarios. Experimental results demonstrate that the proposed approach enhances clustering quality, achieving an Adjusted Rand Index of 0.86 and a Normalized Mutual Information of 0.87, surpassing state-of-the-art baselines such as LogUAD and Log2graphs. Anomaly detection was obtained by the model with an F1-score of 0.91 and an AUC of 0.95, and evidence of stability in detecting deviations at low false alarm rates. Computational efficiency analysis also indicates that inference latency reduces by ~26% compared to graph-based baselines, with the ability to support real-time monitoring. These results establish that the intended approach not only enables more design-time verification of SoC systems but also facilitates secure runtime fault monitoring. The paper concludes that machine learning-based behavioural abstraction is an operationally tractable, interpretable, and scalable solution for enhancing SoC dependability. Subsequent research will deploy the approach to heterogeneous log sources, energy-constrained optimization, and adaptive online learning across changing workloads.

Abstract A central challenge in the control of complex industrial processes like heat exchangers is the persistent disconnection between global offline optimization and real-time online adaptation, leading to suboptimal performance under dynamic conditions. To address this, this paper proposes a coordinated hierarchical control strategy, CB-BPP, a three-layer framework integrating Chaotic Particle Swarm Optimization (CFPSO), a Backpropagation Neural Network (BPNN) identifier, and a Predictive Functional Control (PFC)-based PID controller. At the top layer, CFPSO performs offline global optimization of the BPNN's initial weights and establishes safe bounds for controller parameters. The middle layer employs the BPNN for online system identification and real-time adaptation of PID parameters. The bottom layer executes high-precision control. This tiered architecture decouples offline optimization from online adaptation, enabling synergistic control. Comprehensive simulations demonstrate the superiority of CB-BPP, reducing the Integral of Absolute Error (IAE) by up to 45% and improving recovery speed by 35% compared with advanced baseline methods. Furthermore, extensive robustness analysis against sensor noise and parameter drift, alongside a successful Hardware-in-the-Loop (HIL) implementation, validates its practical applicability. The results confirm the proposed method's high precision, strong robustness, and real-time feasibility, providing an advanced control solution for complex industrial processes.

Active Inference is a closed-loop computational theoretical basis for understanding behaviour, based on agents with internal probabilistic generative models that encode their beliefs about how hidden states in their environment cause their sensations. We review Active Inference and how it could be applied to model the human–computer interaction loop. Active Inference provides a coherent framework for managing generative models of humans, their environments, sensors and interface components. It informs off-line design and supports real-time, online adaptation. It provides model-based explanations for behaviours observed in HCI, and new conceptual tools with the potential to measure important concepts such as agency and engagement in interaction. We discuss how Active Inference offers a new basis for a theory of interaction in HCI, tools for design of modern, complex sensor-based systems, and integration of artificial intelligence technologies, enabling it to cope with diversity in human users and contexts. We discuss the practical challenges in implementing such Active Inference-based systems.

Accurate control of integrating and non-self-regulating processes remains a persistent challenge in process industries due to their inherent open-loop instability, slow response characteristics, and high sensitivity to parametric uncertainties and external disturbances. These characteristics often render conventional proportional-integral-derivative (PID) controllers inadequate, particularly under varying operating conditions or fault scenarios. To address these limitations, this study introduces a novel hybrid PID tuning framework that integrates the global optimization capability of the Harris Hawks Optimization (HHO) algorithm with a neural network-based supervisory model. The neural network is trained to dynamically estimate admissible bounds for key performance indices such as integral absolute error (IAE), integral squared error (ISE), and integral time-weighted absolute error (ITAE) based on process behavior. These bounds serve as adaptive constraints in the optimization phase, allowing the tuning mechanism to remain process-specific and responsive to variations, faults, and non-linearities. The overall controller design is formulated as a constrained multi-objective optimization problem, where the objectives include minimizing control errors while simultaneously satisfying robustness, stability, and performance constraints. Unlike traditional fixed-rule or purely heuristic-based tuning techniques, the proposed approach enables online adaptation to disturbances and faults by leveraging real-time feedback from the neural network. This enhances both robustness and fault tolerance across a wide range of operating scenarios. The effectiveness of the proposed method is rigorously evaluated through extensive simulations on several benchmark integrating processes under both nominal and faulty conditions, including sensor and actuator faults. Comparative analysis with recent methods confirms that the proposed controller offers superior tracking accuracy, faster settling time, and enhanced robustness. In addition, the robustness of the closed-loop system is graphically validated using frequency-domain magnitude plots under multiplicative input and output uncertainties. These results confirm the practical value and innovative nature of the proposed intelligent hybrid tuning strategy.