Testing of e-government services relies on privacy-preserving synthetic test data that are as similar as possible to actual real-life raw data, but also satisfy the requirements of the test cases. Obtaining such test data in Estonian e-government settings is a resource-intensive and largely manual process. This work addresses the challenges of the current process and suggests a novel synthetic test data generation approach that is largely automated, does not require access to real-life raw data, and generates realistic synthetic test data for Estonian e-government settings. We validate the Proof of Concept of our novel synthetic test data generation approach in real-life-like settings. We conclude that the approach can already generate synthetic data for simpler test cases, but needs additional work to cover more complex test cases that require synthetic test data to be compatible across different data services.

Operational hazard analysis of aviation system operations must consider interactions among weather, ATC actions, airspace constraints, aircraft operations, and human factors - distinct from the functional hazard assessment applied at the aircraft-system level. We present an AI-assisted approach that generates candidate hazard scenarios from NASA's Aviation Safety Reporting System (ASRS). Given a target adverse outcome, it produces a structured hypothesis as categorical factors and a narrative scenario describing an operational event sequence consistent with the structure. Each scenario includes by a plausibility score from historical co-occurrence evidence and traceability to the most similar held-out ASRS reports. We then propose a hybrid variant, conditioning narrative generation on a structured hypothesis produced via evolutionary abduction, improving correctness and reducing variability. We evaluate multiple large language models, zero-shot versus few-shot prompting, and optional fine-tuning, measuring how prompting and model choice affect the validity and realism of the generated structures and narratives.

This study investigates the use of generative AI within a university-level software engineering (SE) course by analysing survey responses from second-year students. The survey explored multiple dimensions of students' interaction with AI while working on their course projects, including frequency of use, task types, prompting strategies, and perceived challenges. The results reveal widespread use of AI, especially for coding and debugging, though students also apply it across other phases of the software development lifecycle (SDLC). Despite this broad engagement, and even some use of advanced prompting, understanding of these techniques remains limited. All students reported verifying AI outputs, indicating low trust, which is further reinforced by the fact that many cited the inaccuracy of AI-generated results as their biggest challenge. Students also expressed a clear interest in improving their skills, particularly in prompt design. These findings underscore the need for structured support in AI literacy and prompting skills, as well as adapting course projects for an AI-enhanced learning context. This study provides a foundation for future research and instructional design in SE education.

The manufacturing sector's increasing reliance on Industry 4.0 technologies has made it a prime target for ransomware attacks, which can disrupt operations, cause financial losses, and compromise intellectual property. While prior studies have explored ransomware threats to industrial systems, few have leveraged dark web-disclosed data to understand the scale and nature of these attacks. This study analyzes 7,427 ransomware attack records disclosed on dark web onion services from April 2022 to March 2025, focusing on the manufacturing sector. The dataset, initially comprising 10,000 records, was cleaned by removing duplicates and records with missing NAICS codes, inferred using an AI-based approach. Findings reveal that manufacturing is vulnerable, with 1,620 attacks (21.81% of the total), tied with Professional Services as the most targeted sector. The United States accounted for 49.78% of manufacturing attacks, followed by Germany (7.10%), reflecting their significant manufacturing bases. A diverse set of 88 ransomware groups (78.57% of the total 112) targeted manufacturing, with LockBit responsible for 22.41% of attacks. These results underscore the urgent need for tailored cybersecurity strategies in manufacturing, including enhanced OT security and international collaboration to mitigate ransomware threats, particularly in high-risk regions like the U.S. and Germany.

Background: High-level system testing of applications that use data from e-Government services as input requires test data that is real-life-like but where the privacy of personal information is guaranteed. Applications with such strong requirement include information exchange between countries, medicine, banking, etc. This review aims to synthesize the current state-of-the-practice in this domain. Objectives: The objective of this Systematic Review is to identify existing approaches for creating and evolving synthetic test data without using real-life raw data. Methods: We followed well-known methodologies for conducting systematic literature reviews, including the ones from Kitchenham as well as guidelines for analysing the limitations of our review and its threats to validity. Results: A variety of methods and tools exist for creating privacy-preserving test data. Our search found 1,013 publications in IEEE Xplore, ACM Digital Library, and SCOPUS. We extracted data from 75 of those publications and identified 37 approaches that answer our research question partly. A common prerequisite for using these methods and tools is direct access to real-life data for data anonymization or synthetic test data generation. Nine existing synthetic test data generation approaches were identified that were closest to answering our research question. Nevertheless, further work would be needed to add the ability to evolve synthetic test data to the existing approaches. Conclusions: None of the publications really covered our requirements completely, only partially. Synthetic test data evolution is a field that has not received much attention from researchers but needs to be explored in Digital Government Solutions, especially since new legal regulations are being placed in force in many countries.

Resistance training is used to prevent the decrease in muscle strength associated with age. This study evaluates the effectiveness of iso-inertial training on power, physical performance, and risk of falls compared to gravitational training in physically active middle-older adults. Parallel-group, randomised controlled trial at Espai Esport Wellness Center (Granollers, Spain). Forty-four physically active adults (age >57) were randomised to iso-inertial (n=21) or gravitational (n=23) training groups (R software; 1:1 ratio). Participants had to complete a 6-week training program (2 sessions/week) consisting of three exercises (forward and side lunge, forward lunge with row). Primary outcome: power in the eccentric phase of each exercise evaluated with both iso-inertial and gravitational devices. Secondary outcomes: concentric power, physical performance, risk of falls. Only outcome evaluators were blinded. We used multivariate linear regression models for the analysis. 27 participants completed the program (n=15 iso-inertial, n=12 gravitational). Iso-inertial training showed better eccentric power gains than gravitational training for the iso-inertial system evaluation, although the difference was only statistically significant for the side lunge. Forward lunge: between-group difference 4.50 W (95% CI: -2.94 to 11.94, p=0.23); side lunge: difference 9.24 W (95% CI: 2.99 to 15.49; p=0.00); forward lunge with row: difference 15.25 W (95% CI: -0.63 -to 31.13, p=0.06). We observed no differences for the gravitational system evaluation nor for concentric power, physical performance, and risk of falls. The two groups improved remarkably from baseline for all outcomes. Iso-inertial training leads to better eccentric power gains than gravitational training and therefore could be considered by clinicians when prescribing resistance training to middle-older adults. Both training systems were equally effective in improving concentric power and physical performance, and reducing risk of falls. This study was registered at Clinicaltrials.gov (NCT06160089).

The research area of Software Defect Prediction (SDP) is both extensive and popular, and is often treated as a classification problem. Improvements in classification, pre-processing and tuning techniques, (together with many factors which can influence model performance) have encouraged this trend. However, no matter the effort in these areas, it seems that there is a ceiling in the performance of the classification models used in SDP. In this paper, the issue of classifier performance is analysed from the perspective of data complexity. Specifically, data complexity metrics are calculated using the Unified Bug Dataset, a collection of well-known SDP datasets, and then checked for correlation with the defect prediction performance of machine learning classifiers (in particular, the classifiers C5.0, Naive Bayes, Artificial Neural Networks, Random Forests, and Support Vector Machines). In this work, different domains of competence and incompetence are identified for the classifiers. Similarities and differences between the classifiers and the performance metrics are found and the Unified Bug Dataset is analysed from the perspective of data complexity. We found that certain classifiers work best in certain situations and that all data complexity metrics can be problematic, although certain classifiers did excel in some situations.

The COVID-19 pandemic has evolved the way that education takes place. Distant or hybrid learning has confirmed the importance of cloud computing and network infrastructure for maintaining education activities in this situation. However, lessons learned from these experiences also show the problems of education exclusion and digital infrastructure limitations. In this paper, we analyze these two problems and propose a solution named POEMA, a Personal Cloud for inclusive education based on the cloud continuum concepts and beyond with explicit consideration of inclusiveness workloads by design. Likewise we propose an Inclusiveness Education Key Value Indicator (KVI) definition compatible with the 6G Key Value Indicator concept of Hexa-X.

ABSTRACT Introduction: The objective of this study was to present a systematic review and meta-analysis to compare total excess post-exercise oxygen consumption (EPOC) for two training intervention models in healthy individuals, and the secondary objective was to understand whether oxygen consumption after exercise could really promote a meaningful help. Design: To design a meta-analysis review to compare two training intervention models (experimental: high-intensity interval training; and control: continuous moderate-intensity) and their effects on total EPOC in healthy individuals. Participants: Seventeen studies were considered to be of good methodological quality and with a low risk of bias. Methods: Literature searches were performed using the electronic databases with no restriction on year of publication. The keywords used were obtained by consulting Mesh Terms (PubMed) and DeCS (BIREME Health Science Descriptors). Results: The present study findings showed a tendency (random-effects model: 0.87, 95%-CI [0.35,1.38], I2=73%, p<0.01) to increase EPOC when measured following high-intensity interval training. Conclusions: Our study focused on the analysis of high- and moderate-intensity oxygen uptake results following exercise. Despite the growing popularity of high-intensity interval training, we found that the acute and chronic benefits remain limited. We understand that the lack of a standard protocol and standard training variables provides limited consensus to determine the magnitude of the EPOC. We suggest that longitudinal experimental studies may provide more robust conclusions. Another confounding factor in the studies investigated was the magnitude (time in minutes) of VO2 measurements when assessing EPOC. Measurement times ranged from 60 min to 720 min. Longitudinal studies and controlled experimental designs would facilitate more precise measurements and correct subject numbers would provide accurate effect sizes. Systematic reviewb of Level II studies.

The EU research project between industry and academia mu DevOps is a collaborative research project formed by an international network of organizations including industry and academia that aims to tackle current challenges of microservice development operations. An important case study considered in this project is the Cyber Ranges application a cyber security training and capability development exercises using microservices for the design, delivery, and management of simulation-based, experiences in cyber security developed by Silensec as one of the partners. This work describes the results of analyzing the scenario usage dataset of the Cyber Ranges training platform. This includes the matrix of starts for scenario/user and the attributes of scenarios. The aims are to produce recommendations of scenarios for users based on previous activity and to predict the success of scenarios as measured by the number of starts.

Microservice architectures are becoming increasingly important since they facilitate agile and modular production cycles to deliver applications using collections of loosely coupled and fine-grained services. The mu DevOps is a research project formed by an international network of organizations including industry and academia that aims to tackle current challenges of microservice development operations. This paper presents the mu DevOps project and the initial research carried out to evaluate the user experience of a microservice web application that delivers cybersecurity learning. Results point to critical elements of three main functionalities: library of scenarios, scenario information and entering scenario. Since there are several currently available solutions that offer similar services, the user experience of the microservice web app may play a critical role in determining which application will get a dominant role in the market.

A prática de atividade fí­sica auxiliada por uma alimentação equilibrada trazem resultados satisfatórios, mas fatores como a falta de informação, orientação ou recurso financeiro, somados ao anseio de atingir os resultados rapidamente, fazem com que os indiví­duos busquem os recursos ergogênicos. O presente estudo teve como objetivo realizar uma revisão de literatura a respeito do consumo de suplementos alimentares e esteroides (recursos ergogênicos) em academias do Brasil. Foram utilizadas as bases de dados eletrônicas Scielo, Revista Brasileira de Fisiologia do Exercí­cio (RBFE) e Revista Brasileira de Nutrição Esportiva (RBNE), e as seguintes palavras-chave: anabolizantes, esteroides anabólicos androgênicos, suplementos alimentares, recursos ergogênicos, consumo, praticantes de musculação. Foram selecionados 29 artigos. Verificou-se que de 4.877 pessoas praticantes de academia, 40,8% consomem recursos ergogênicos com finalidade estética. Na maioria dos estudos, a indicação para o consumo dos recursos ergogênicos foi de profissionais não habilitados. Apesar das leis de restrição para o comércio dos recursos ergogênicos, visando í proteção dos consumidores, percebe-se que os usuários conseguem ter acesso aos recursos sem a devida orientação profissional, colocando-os em risco. Necessita-se de novas medidas para que haja maior controle sobre a comercialização dos recursos ergogênicos.Palavras-chave: esteroides anabólicos androgênicos, suplementos nutricionais, recursos ergogênicos, praticantes de musculação.

Semi-supervised learning, using both labelled and unlabelled data can be helpful in situations where a large amount of unlabelled data is readily available and labelled data is expensive to create, such as with app review classification.Although semi-supervised learning has been proven to provide meaningful results for a variety of applications, it has not been heavily researched in the field of app review analysis. This research analysed the suitability of semi-supervised classification through self-training when a low proportion of data is labelled. It also experimented with how adding metadata changes the performance of semi-supervised classification and analysed a large dataset of reviews to find how the differences between free and paid apps, app categories, apps with different popularities and the version releases of apps.In this paper we assess the viability of using semi-supervised classification to classify app reviews by analysing a large dataset of reviews extracted from the Apple App Store. We classify the data as either \textit{bugs}, \textit{feature requests} or \textit{other}. We found that semi-supervised learning can be effective even when a low proportion of data is labelled and can perform similarly to a base classifier with 100% labelled data.

Many Pareto-based multiobjective evolutionary algorithms require ranking the solutions of the population in each iteration according to the dominance principle, which can become a costly operation particularly in the case of dealing with many-objective optimization problems. In this article, we present a new efficient algorithm for computing the nondominated sorting procedure, called merge nondominated sorting (MNDS), which has a best computational complexity of O(NlogN) and a worst computational complexity of O(MN2) , with N being the population size and M being the number of objectives. Our approach is based on the computation of the dominance set, that is, for each solution, the set of solutions that dominate it, by taking advantage of the characteristics of the merge sort algorithm. We compare MNDS against six well-known techniques that can be considered as the state-of-the-art. The results indicate that the MNDS algorithm outperforms the other techniques in terms of the number of comparisons as well as the total running time.

Welcome to the Eigth International Workshop on Artificial Intelligence and Requirements Engineering (AIRE’21). The purpose of this workshop is to explore synergies between Artificial Intelligence (AI) and Requirements Engineering (RE). AIRE aims to strengthen the links in the community, including those between industry and academia. As such, we welcome submissions in the intersection between RE and AI. An important goal is to inspire a new and broad community for interdisciplinary discussions concerning novel research directions for Requirements Engineering and Artificial Intelligence. The edition of the workshop in 2021 received 18 submissions, which were independently reviewed by at least three program committee members. In the end, 12 long papers were accepted. All the conflicts of interest were treated seriously and independently. The high quality of submissions are a sign of a healthy research community and the selected papers will lead to a stimulating program, which also includes technical presentations and the Panel. We hope that you enjoy the AIRE’21 workshop and its proceedings. We consider that in the days when AI is gaining prominence in our daily lives, the RE community cannot neglect the benefit that AI techniques can deliver to the practice of requirements engineering. We look forward to seeing you all at this workshop and the future editions. We are very grateful to the Program Committee members and authors of the submissions for their hard work and dedication in putting together this program. We would like to thank you all for your participation in AIRE’21. We hope that you find this workshop fruitful and inspiring!Nelly, Rachel, and Daniel

ABSTRACT Introduction: The benefits of strength training (ST) include not only strength improvement but also favorable body composition changes, which has led to a considerable increase in the indication of this training method in overweight and obese individuals, and has made the investigation of outcomes attributed to different manipulations of ST variables an important task. However, acute metabolic responses related to energy expenditure (EE) associated with the manipulation of exercises which, in turn, are associated with the number of joints involved in movement, are still inconclusive. Objective: To verify the influence of the number of joints involved in movement on EE with equalized volume in ST at different intensities. Methods: This training program was held on alternate days, with a 48-hour interval between each session, and with two randomized protocols, as follows: multi joint protocol with four common exercises for ST participants compared to the single joint protocol with four exercises. Each protocol was evaluated at three training intensities (90%, 75% and 60% of 1-RM) according to the one-repetition maximum test. Results: Significant increases in EE were observed in the multi joint session as compared to the single joint session: 90% 1-RM multi joint 246.80 ± 26.17 kcal vs single joint 227.40 ± 24.54 kcal (∆ -7.86, 95% CI 7.33; 31.46; t 3.44; p <0.05); 75% 1-RM multi joint 124.13 ± 25.40 kcal vs single joint 111.80 ± 22.78 kcal (∆ -9.93, 95% CI 3.25; 21.41; t 2.91; p <0.05); 60% 1-RM multi joint 70.80 ± 6.28 kcal vs single joint 64.40 ± 6.72 kcal (∆-9.04, 95% CI 3.95; 8.84; t 5.60; p <0.05). Conclusion: Multi joint exercises may be a variable to consider when EE balance is the main target of the ST program. However, further studies are needed to supplement our findings. Level of evidence II; Diagnostic studies-Investigating a diagnostic test.

Abstract An accurate method for quantifying associated metabolic cost has yet to be developed for a strength training session (ST). The aim of this study was to quantify the energy expenditure (EE) in an ST session composed of eight exercises at moderate intensity using indirect calorimetry and, from the values obtained, develop a prediction equation for estimating EE. Fifteen males (22.9 ± 2.61 years old), with at least 12 months of experience in ST performed one session of strength training composed of 8 exercises. Three sets of repetitions were performed until concentric failure for each exercise at 75% of 1-repetition maximum (75% of 1RM). The model demonstrated that session time and load volume of ST was a significant predictor of EE (p < 0.05). We found that the energy cost of an ST session at an intensity of 75% of 1RM could be predicted using the equation of Y’ = −473.595 + −1.2110(X1) + 17.5723(X2) (R2 = 0.61, p < 0.05). Where X1 = load-volume (no. of sets x no. of repetitions); X2 = session time (minutes). Although our equation may have limited accuracy, our regression formula accounted for 61% of the variability in a strength training session at a moderate intensity of 75% of 1RM. Session time in the total variability of EE in ST was an important consideration.