Cowpea and bambara groundnut flours improve amino acid profiles of gruels but often taste bitter to infants. Controlled lactic fermentation with selected starter cultures reduces tannin activity and beany notes while increasing protein digestibility corrected amino acid score. Sensory trials with caregivers and acceptability tests in 120 children aged 6–23 months show parity with commercial fortified cereals at half the ingredient cost.

Manufacturing plants increasingly bridge Modbus RTU devices to cloud analytics through edge gateways with limited RAM and no hardware crypto accelerators. We implement a lattice-based key encapsulation profile tuned for 32-bit microcontrollers and measure handshake latency on three commercial gateways. The protocol completes within existing scan-cycle gaps on production lines while resisting harvest-now-decrypt-later threats posed by long-lived sensor credentials.

Traditional otter surveys are labor-intensive in steep catchments recovering from legacy mining. Passive acoustic loggers deployed at 24 remediated and reference sites capture spraint-associated vocalizations and foraging splashes. Random forest classifiers trained on annotated clips discriminate activity levels that correlate with macroinvertebrate recovery indices, offering a scalable complement to electrofishing and sign surveys for post-remediation monitoring.

Coastal municipalities face chloride ingress while landfills accumulate cullet with limited reuse pathways. Mortar and slab specimens using 20% ground glass by binder mass in seawater-mixed formulations are tested for strength, shrinkage, and chloride penetration over 180 days. Optimal fineness yields comparable 28-day compressive strength to control mixes while lowering embodied carbon, suggesting viability for non-structural promenade applications.

Anticipatory social protection is debated without clear evidence on whether advance transfers outperform reactive payments when staple prices spike. Using four waves of urban household panels linked to market price indices, we estimate dynamic treatment effects of a municipal top-up program. Households receiving transfers two weeks before peak inflation reduce asset sales and school absenteeism relative to those paid after prices normalize, supporting indexed pre-positioning.

Photovoltaic pumps stall during cloudy harvest weeks when crops need water most. We integrate a compact phase-change thermal store that drives a scroll pump after sunset using daytime collector surplus. Two-year measurements on three cooperative farms show 28% higher seasonal water delivery compared with battery-only systems of equal capital cost, with lower maintenance from reduced deep-cycling of lithium packs.

Screening backlogs leave many diabetic patients without timely ophthalmology referral. We fine-tune a vision transformer on 18,000 fundus photographs collected through a low-cost smartphone adapter deployed in six district clinics. The model achieves sensitivity above 0.91 for referable macular edema while maintaining specificity suitable for triage, reducing unnecessary specialist visits by roughly one third in a six-month operational pilot.

Carfilzomib is a second-generation proteasome inhibitor widely used in the treatment of multiple myeloma, while its neurological adverse effects remain clinically relevant and may affect treatment tolerance and patient quality of life. However, there is still a lack of efficient in vivo platforms for integrated evaluation of carfilzomib-associated functional neural impairment and preliminary screening of potential protective compounds. In this study, we established a zebrafish larval model to characterize carfilzomib-induced neurobehavioral impairment and to assess its applicability for compound screening. Zebrafish larvae were exposed to graded concentrations of carfilzomib until 144 hpf. Locomotor behavior was quantified using an automated tracking system, and molecular alterations were evaluated by assessing neuronal function-related gene expression (syn2α, gap43, dat, bdnf, and gdnf) together with ROS levels. Carfilzomib exposure induced a concentration-dependent reduction in locomotor activity, accompanied by coordinated downregulation of neuronal function-related genes and increased ROS accumulation, indicating coupled behavioral and molecular dysregulation. To evaluate the screening potential of this model, acetyl-L-carnitine and α-lipoic acid were tested as representative candidate neuroprotective compounds. Both compounds significantly alleviated locomotor impairment, partially restored neuronal gene expression, and reduced oxidative stress. Collectively, these findings demonstrate that carfilzomib induces a reproducible neurobehavioral impairment phenotype in zebrafish larvae characterized by synchronized locomotor deficits, neuronal function-related gene dysregulation, and oxidative stress. This integrated model provides a validated and efficient in vivo platform for functional assessment of chemotherapy-associated neural impairment and for preliminary screening of candidate neuroprotective compounds.

Micromanagement remains a contentious leadership approach within higher education institutions, particularly in relation to its perceived role in ensuring employee accountability. This study explored the relationship between micromanagement practices and employee accountability, examining whether close supervision and control mechanisms effectively promote responsible work behaviour among administrative staff at a selected University of Technology (UoT). Drawing on contemporary leadership theories and organizational behaviour frameworks, the study investigates both the potential benefits and drawbacks of micromanagement in complex institutional environments. Using a qualitative research methodology, data was collected from administrative staff across various departments at a selected University of Technology (UoT) to examine perceptions of micromanagement, levels of autonomy, job satisfaction, and accountability outcomes. The findings suggest that while micromanagement may, in some cases, lead to short-term improvements in task completion and compliance, it often undermines employee morale, reduces intrinsic motivation, and limits professional autonomy, factors that are critical for sustained accountability and performance in academic settings. The study further highlights the importance of balanced leadership approaches that integrate trust, clear expectations, and supportive supervision rather than excessive control. It concludes that micromanagement is neither the only nor the most effective strategy for fostering accountability in higher education. Instead, institutions should adopt adaptive leadership styles that prioritize engagement, empowerment, and mutual accountability. These findings contribute to the ongoing discourse on leadership effectiveness in higher education institutions and provide practical insights for managers seeking to enhance accountability without compromising employee well-being and institutional performance.

This paper develops the Brighton Mahohoho Assembled Actuarial Model, an integrated framework for automating core non-life insurance functions, namely actuarial loss reserving, actuarial pricing, and actuarial underwriting, through artificial intelligence-driven data analytics. The proposed model further consolidates four major lines of business, specifically microfinance services, motor insurance, property insurance, and an all-in-one bundled insurance platform, into a unified actuarial decision-support system. The framework is designed to enhance operational efficiency by mitigating the prevalence, incidence, frequency, and severity of insurance claims, while simultaneously reducing unpaid premiums and improving policyholder retention through improved risk segmentation and product integration. The frequency-severity modelling structure is extended using ten machine learning algorithms to construct an Automated Actuarial Loss Reserving, Risk Pricing, and Underwriting Model, further adjusted for inflation effects. Empirical evaluation demonstrates strong predictive performance across all models considered. Among the competing algorithms, the Least Angle Regression (LAR) model exhibits superior predictive accuracy and stability, and is subsequently adopted as the core engine for developing actuarial underwriting structures across the defined business lines. The findings highlight the effectiveness of the Brighton Mahohoho actuarial assembly approach in advancing data-driven, inflation-adjusted non-life insurance modelling under modern actuarial and machine learning paradigms.

This study proposes an advanced actuarial framework for marine insurance risk pricing through\nthe application of Generalized Boosting Machines (GBM). The proposed model leverages ensemble\nlearning techniques to capture complex, non-linear relationships inherent in marine insurance data,\nenhancing predictive accuracy for both risk estimation and premium determination. By integrating\nheterogeneous data sources and embedding IFRS 17 regulatory principles, the framework ensures\nboth actuarial robustness and regulatory compliance in premium formulation. The GBM-based architecture\nis designed to improve loss prediction, risk segmentation, and pricing adequacy under volatile\nmaritime risk environments. Empirical findings demonstrate that the proposed model outperforms\ntraditional actuarial pricing approaches in terms of predictive stability, bias reduction, and overall robustness.\nThese results highlight the potential of machine learning-driven actuarial systems in modern\nnon-life insurance pricing, particularly within the complex domain of marine risk.

This paper introduces the Brighton Mahohoho Automated IFRS17–Solvency II–NASA Protocol Assembled Astro-Actuarial Risk Modelling and Engineering Synergies Framework, a novel interdisciplinary approach that integrates actuarial science, space physics, cosmochemistry, and intelligent robotic systems for the purpose of resilient space mission design. By aligning insurance-based risk quantification principles under IFRS17 and Solvency II with NASA’s mission assurance protocols, the framework establishes a unified methodology for modelling frequency–severity losses, capital adequacy, and dynamic reserves in environments characterized by extreme uncertainty. The methodology incorporates advanced non-linear regression, Gaussian process modelling, and intelligent automation to produce robust estimates of astro-actuarial loss reserves and risk premiums. The results demonstrate not only compliance with financial reporting and solvency requirements but also their adaptability to cosmic-scale hazards, mission uncertainties, and interplanetary operational risks. The framework thereby provides a pioneering blueprint for advancing both actuarial practice and aerospace risk engineering.

Spacecraft launch failures represent rare but high-consequence events with profound financial and operational risks for stakeholders in the aerospace industry. Traditional actuarial models often struggle to capture the complexity, uncertainty, and non-linear relationships inherent in launch failure data due to sparse observations and high variability. This paper introduces an integrated actuarial risk modelling framework that combines Bayesian inference with Gaussian Process Regression (GPR) to assess and predict the probability and severity of spacecraft launch failures. The Bayesian approach allows for probabilistic reasoning under uncertainty and incorporates expert knowledge, while GPR captures non-linear dependencies and complex relationships within sparse datasets. This dual-modelling approach enhances risk quantification accuracy, improves underwriting decisions, and supports IFRS 17 and Solvency II compliance. The paper demonstrates the effectiveness of this framework using simulated and real-world launch data, providing detailed risk metrics, posterior distributions, and predictive uncertainty estimates. The findings offer significant advancements in actuarial risk modelling, with direct applications for insurers, reinsurers, and risk managers in the rapidly evolving space industry.

This study evaluates a school-based educational model that integrates catfish farming with digital learning to promote sustainable education and lifelong learning among Malaysian high school students. Unlike previous studies that have examined aquaculture education, digital learning, or entrepreneurship independently, this study evaluates their holistic integration within a single experiential learning framework. Using a quantitative research design, survey data were collected from 72 Malaysian high school students who participated in the program to evaluate learning outcomes related to aquaculture competencies, digital literacy, technological proficiency, and entrepreneurial development. The findings revealed significant improvements in students' practical aquaculture skills, digital competencies, and entrepreneurial understanding while fostering greater interest in agriculture-related careers. Beyond individual learning outcomes, the integrated model contributed to school income generation, strengthened community engagement through local fish distribution, and advanced sustainable development initiatives. The study demonstrates that integrating school-based catfish farming with digital learning offers a scalable and replicable educational model that promotes experiential learning, digital literacy, entrepreneurship, and lifelong learning while contributing to the Sustainable Development Goals (SDGs). These findings provide practical and theoretical insights for educators, policymakers, and educational institutions seeking sustainable, technology-enhanced educational approaches to preparing students for future technological, environmental, and socio-economic challenges.