The rapid advance of Generative AI into software development prompts this empirical investigation of perceptual effects on practice. We study the usage patterns of 147 professional developers, examining perceived correlates of AI tools use, the resulting productivity and quality outcomes, and developer readiness for emerging AI-enhanced development. We describe a virtuous adoption cycle where frequent and broad AI tools use are the strongest correlates of both Perceived Productivity (PP) and quality, with frequency strongest. The study finds no perceptual support for the Quality Paradox and shows that PP is positively correlated with Perceived Code Quality (PQ) improvement. Developers thus report both productivity and quality gains. High current usage, breadth of application, frequent use of AI tools for testing, and ease of use correlate strongly with future intended adoption, though security concerns remain a moderate and statistically significant barrier to adoption. Moreover, AI testing tools' adoption lags that of coding tools, opening a Testing Gap. We identify three developer archetypes (Enthusiasts, Pragmatists, Cautious) that align with an innovation diffusion process wherein the virtuous adoption cycle serves as the individual engine of progression. Our findings reveal that organizational adoption of AI tools follows such a process: Enthusiasts push ahead with tools, creating organizational success that converts Pragmatists. The Cautious are held in organizational stasis: without early adopter examples, they don't enter the virtuous adoption cycle, never accumulate the usage frequency that drives intent, and never attain high efficacy. Policy itself does not predict individuals' intent to increase usage but functions as a marker of maturity, formalizing the successful diffusion of adoption by Enthusiasts while acting as a gateway that the Cautious group has yet to reach.
With the advancement of Agentic AI, researchers are increasingly leveraging autonomous agents to address challenges in software engineering (SE). However, the large language models (LLMs) that underpin these agents often function as black boxes, making it difficult to justify the superiority of Agentic AI approaches over baselines. Furthermore, missing information in the evaluation design description frequently renders the reproduction of results infeasible. To synthesize current evaluation practices for Agentic AI in SE, this study analyzes 18 papers on the topic, published or accepted by ICSE 2026, ICSE 2025, FSE 2025, ASE 2025, and ISSTA 2025. The analysis identifies prevailing approaches and their limitations in evaluating Agentic AI for SE, both in current research and potential future studies. To address these shortcomings, this position paper proposes a set of guidelines and recommendations designed to empower reproducible, explainable, and effective evaluations of Agentic AI in software engineering. In particular, we recommend that Agentic AI researchers make their Thought-Action-Result (TAR) trajectories and LLM interaction data, or summarized versions of these artifacts, publicly accessible. Doing so will enable subsequent studies to more effectively analyze the strengths and weaknesses of different Agentic AI approaches. To demonstrate the feasibility of such comparisons, we present a proof-of-concept case study that illustrates how TAR trajectories can support systematic analysis across approaches.
To accurately evaluate the dynamic characteristics of helical gear transmission systems under high-speed operating conditions and to address the problem that the lumped parameter model fails to accurately reflect the system’s dynamic performance because its degrees of freedom are imperfectly considered, the full-degree-of-freedom dynamic model of the high-speed shaft-bearing-helical gear transmission system was established. The system’s dynamic equations were derived, and the natural frequencies calculated from the theoretical model, as well as the corresponding mode shapes for each degree of freedom, were compared with the finite element model simulation results. The relationship between the critical rotational speed within the system’s 3 × 104 rpm range of rotational speed and the system frequency was analyzed. The dynamic characteristics of the driving wheel and driven wheel in terms of amplitude and motion trajectory were investigated, and the influence of eccentricity on these characteristics was discussed. Results indicate that the relative deviation of the first six natural frequencies between the proposed model and finite element model ranges from 0.09 % to 13.1 %, and the deviations of the higher-order modes are within 1 %. This effectively validates that the full-degree-of-freedom model can accurately reflect the dynamic characteristics of the system, providing a theoretical basis for the design of high-speed-helical gear transmission system.
Applications of Large Language Models (LLMs) are rapidly growing in industry and academia for various software engineering (SE) tasks. As these models become more integral to critical processes, ensuring their reliability and trustworthiness becomes essential. Consequently, the concept of trust in these systems is becoming increasingly critical. Well-calibrated trust is important, as excessive trust can lead to security vulnerabilities, and risks, while insufficient trust can hinder innovation. However, the landscape of trust-related concepts in LLMs in SE is relatively unclear, with concepts such as trust, distrust, and trustworthiness lacking clear conceptualizations in the SE community. To bring clarity to the current research status and identify opportunities for future work, we conducted a comprehensive review of $88$ papers: a systematic literature review of $18$ papers focused on LLMs in SE, complemented by an analysis of 70 papers from broader trust literature. Additionally, we conducted a survey study with 25 domain experts to gain insights into practitioners' understanding of trust and identify gaps between existing literature and developers' perceptions. The result of our analysis serves as a roadmap that covers trust-related concepts in LLMs in SE and highlights areas for future exploration.
Mohammed Latif Siddiq, Arvin Islam-Gomes, Natalie Sekerak
et al.
Reproducibility is a cornerstone of scientific progress, yet its state in large language model (LLM)-based software engineering (SE) research remains poorly understood. This paper presents the first large-scale, empirical study of reproducibility practices in LLM-for-SE research. We systematically mined and analyzed 640 papers published between 2017 and 2025 across premier software engineering, machine learning, and natural language processing venues, extracting structured metadata from publications, repositories, and documentation. Guided by four research questions, we examine (i) the prevalence of reproducibility smells, (ii) how reproducibility has evolved over time, (iii) whether artifact evaluation badges reliably reflect reproducibility quality, and (iv) how publication venues influence transparency practices. Using a taxonomy of seven smell categories: Code and Execution, Data, Documentation, Environment and Tooling, Versioning, Model, and Access and Legal, we manually annotated all papers and associated artifacts. Our analysis reveals persistent gaps in artifact availability, environment specification, versioning rigor, and documentation clarity, despite modest improvements in recent years and increased adoption of artifact evaluation processes at top SE venues. Notably, we find that badges often signal artifact presence but do not consistently guarantee execution fidelity or long-term reproducibility. Motivated by these findings, we provide actionable recommendations to mitigate reproducibility smells and introduce a Reproducibility Maturity Model (RMM) to move beyond binary artifact certification toward multi-dimensional, progressive evaluation of reproducibility rigor.
Zaid Ameen Abduljabbar, Vincent Omollo Nyangaresi, Ahmed Ali Ahmed
et al.
Abstract Vehicular Ad-Hoc Networks (VANETs) have facilitated the massive exchange of real-time traffic and weather conditions, which have helped prevent collisions, reduce accidents, and road congestions. This can effectively enhance driving safety and efficiency in technology-driven transportation systems. However, the transmission of massive and sensitive information across public wireless communication channels exposes the transmitted data to a myriad of privacy as well as security threats. Although past researches has developed many vehicular ad-hoc networks security preservation schemes, several of them are inefficient or susceptible to attacks. This work, introduces an approach that leverages reverse fuzzy extraction, bilinear pairing, and Physically Unclonable Function (PUF) to design an efficient and anonymity-preserving authentication scheme. We conduct an elaborate formal security analysis to demonstrate that the derived session key is secure. The semantic security analyses also demonstrate its resilience against typical VANET attacks such as impersonations, denial of service, and de-synchronization, instilling confidence in its effectiveness. Moreover, our approach incurs the lowest computational overheads at relatively low communication costs. Specifically, our protocol attains a 66.696% reduction in computation costs, and a 70% increment in the supported security functionalities.
In internet service industries, such as competitive industries, it costs more to attract new consumers to become customers of the company than saving the consumers who already are customers. Therefore, detecting the running off customers and finding a way to keep the customers from leaving is extremely important. This study addresses the problem of customer attrition in the internet service industry by choosing the best-performing model to detect the customers who are going to run off in advance. To select the most suitable model for accurately detecting customer churn, this study performs preprocessing, including data cleaning, feature engineering, and feature selection. The dataset is then split into training, testing, and validation sets. Various models are built and evaluated based on their performance, measured by calculating the mean and standardized values of the detection rate. The result is that the Light Gradient Boosting Machine (LGBM) model has superior performance in detection rate scoring.
Paulo Roger Lopes Alves, German Andres Estrada-Bonilla, Antonio Marcos Miranda Silva
et al.
This study explored how different sugarcane vinasses influence the structure and composition of soil bacterial communities in two tropical Oxisols with contrasting textures. In a controlled microcosm experiment with sugarcane seedlings, two concentrations of three vinasse types were applied, and bacterial communities were monitored over 10, 30, and 60 days using T-RFLP and 16S rRNA gene sequencing. Across all treatments, vinasse application led to clear changes in bacterial community structure in both soils, regardless of the time point. Certain bacterial groups, such as <i>Sphingobacteriia</i>, <i>Alphaproteobacteria</i>, and <i>Gammaproteobacteria</i>, became more abundant—likely responding to increased carbon availability, higher pH, and greater soil moisture. At the same time, other groups declined, possibly due to excess nutrients like potassium and sulfur. Notably, these shifts occurred even when standard biochemical indicators suggested no major impact, highlighting the sensitivity of microbial community-level responses. These findings point to the importance of looking beyond traditional soil quality metrics when assessing the environmental effects of organic residue applications. Incorporating microbial indicators can offer a more nuanced understanding of how practices like vinasse reuse affect soil functioning in tropical agroecosystems.
Bilingual teaching resources are insufficient in applied universities across the Southwest China. This study constructs a Structural Equation Model (SEM) based on four latent variables: student factors, teacher factors, external factors, and teaching effects. Data were collected through a questionnaire administered to 550 undergraduates majoring in specific disciplines of architecture and civil engineering at applied universities in Southwest China. Quantitative analysis yielded the following key findings: (1) external factors have a direct and positive influence on both teacher and student variables, with teaching resources exerting the strongest effect among external factors; (2) teacher factors positively affect student factors, with teachers’ attitudes playing the most critical role; and (3) both teacher and student factors significantly impact bilingual teaching effectiveness, with student quality being the most influential component among student-related variables. By integrating external, teacher, and student dimensions, the study proposes targeted strategies to improve bilingual education outcomes. The study provides new insights into the key determinants of bilingual teaching effectiveness and fills a research gap by applying SEM to quantitatively analyze bilingual education in the context of applied universities. It also offers valuable implications for educational administrators and government policymakers seeking to enhance the quality of bilingual education in Southwest China.
The eastern parts of India, Bangladesh, and the North Eastern Region (NER) of India are among the world's most active areas for thunderstorms and lightning, resulting in substantial human and livestock casualties annually. Lightning strikes cause over 500 fatalities in India and approximately 300 in Bangladesh each year, alongside significant property damage and loss of wildlife, such as the death of 18 elephants in Assam in 2021. To mitigate the devastating effects, this study developed a pilot-scale forecasting system for lightning and thunderstorms over the NER of India, leveraging space-based platforms, ground-based detectors, and numerical models. Data sources included the WWLLN (World Wide Lightning Location Network), India's national lightning detection network, Doppler Weather Radar (DWR), and satellite data from INSAT 3D/3DS. The WRF-ELEC model was employed for forecasting, assimilating lightning data via nudging techniques, and achieving forecasts with up to 75% accuracy for lead times of up to four hours. A GIS-based system was used to track convective systems and predict impacted areas at a village level with a one-hour lead time. This system integrates lightning detection, satellite imagery, and DWR data, enabling the identification of affected populations and land use, thereby aiding in disaster preparedness and mitigation. The study demonstrates the potential of integrating earth observation data, in-situ measurements, and numerical models to provide location-specific and time-sensitive lightning forecasts. Coupled with awareness campaigns on safety measures during lightning, this approach offers a robust mechanism to reduce casualties and property damage. Scaling this system beyond the pilot region could significantly enhance disaster risk reduction in other lightning-prone regions.
Ebube Alor, Ahmad Abdellatif, SayedHassan Khatoonabadi
et al.
Software engineering (SE) chatbots are increasingly gaining attention for their role in enhancing development processes. At the core of chatbots are Natural Language Understanding platforms (NLUs), which enable them to comprehend user queries but require labeled data for training. However, acquiring such labeled data for SE chatbots is challenging due to the scarcity of high-quality datasets, as training requires specialized vocabulary and phrases not found in typical language datasets. Consequently, developers often resort to manually annotating user queries -- a time-consuming and resource-intensive process. Previous approaches require human intervention to generate rules, called labeling functions (LFs), that categorize queries based on specific patterns. To address this issue, we propose an approach to automatically generate LFs by extracting patterns from labeled user queries. We evaluate our approach on four SE datasets and measure performance improvement from training NLUs on queries labeled by the generated LFs. The generated LFs effectively label data with AUC scores up to 85.3% and NLU performance improvements up to 27.2%. Furthermore, our results show that the number of LFs affects labeling performance. We believe that our approach can save time and resources in labeling users' queries, allowing practitioners to focus on core chatbot functionalities rather than manually labeling queries.
Due to the limitations of small targets in remote sensing images, such as background noise, poor information, and so on, the results of commonly used detection algorithms in small target detection is not satisfactory. To improve the accuracy of detection results, we develop an improved algorithm based on YOLOv8, called LAR-YOLOv8. First, in the feature extraction network, the local module is enhanced by using the dual-branch architecture attention mechanism, while the vision transformer block is used to maximize the representation of the feature map. Second, an attention-guided bidirectional feature pyramid network is designed to generate more discriminative information by efficiently extracting feature from the shallow network through a dynamic sparse attention mechanism, and adding top–down paths to guide the subsequent network modules for feature fusion. Finally, the RIOU loss function is proposed to avoid the failure of the loss function and improve the shape consistency between the predicted and ground-truth box. Experimental results on NWPU VHR-10, RSOD, and CARPK datasets verify that LAR-YOLOv8 achieves satisfactory results in terms of mAP (small), mAP, model parameters, and FPS, and can prove that our modifications made to the original YOLOv8 model are effective.
Elizabeth Bjarnason, Mirko Morandini, Markus Borg
et al.
The RET (Requirements Engineering and Testing) workshop series provides a meeting point for researchers and practitioners from the two separate fields of Requirements Engineering (RE) and Testing. The goal is to improve the connection and alignment of these two areas through an exchange of ideas, challenges, practices, experiences and results. The long term aim is to build a community and a body of knowledge within the intersection of RE and Testing, i.e. RET. The 2nd workshop was held in co-location with ICSE 2015 in Florence, Italy. The workshop continued in the same interactive vein as the 1st one and included a keynote, paper presentations with ample time for discussions, and a group exercise. For true impact and relevance this cross-cutting area requires contribution from both RE and Testing, and from both researchers and practitioners. A range of papers were presented from short experience papers to full research papers that cover connections between the two fields. One of the main outputs of the 2nd workshop was a categorization of the presented workshop papers according to an initial definition of the area of RET which identifies the aspects RE, Testing and coordination effect.
Tatyana D. Kartuzova, Natalya A. Shelestova, Alexander A. Tkachev
et al.
Purpose: to solve the problem of conservation and reproduction of fish stocks of valuable commercial fish species in natural reservoirs, on the example of performing one of the tasks of developing a facility for regulating water regime of the ilmen part of the Volga delta. Using the well-known formulas of hydraulics, scientific developments of scientists in the field of hydraulic engineering and a priori information, it is necessary to determine the main parameters of a control structure located on a fishery reservoir in Astrakhan region.
Materials and methods: a stage-by-stage hydraulic calculation of a tubular control structure that performs the functions of a water outlet was performed: the carrying capacity of the facility during operation in a pressure mode was determined; the main loads influencing the structure underground contour were determined by filtration calculation; a strength calculation of the floor for shear was performed.
Results. A facility to regulate the water regime of the ilmen area of the Volga delta, ilmen, which are places of natural spawning grounds for commercial fish in Astrakhan region has been developed. The task of hydraulic calculation is to determine the conditions for the advancement of the filtration flow under the structure determining the pressure at characteristic points and to define the minimum dimensions of the floor that perceive stably the loads created by this flow.
Conclusions. The calculation determined the main dimensions of the floor (length, width of the structure – 46.0; 4.65 m; length, thickness of the apron – 4.54; 0.3 m; length, thickness of the water apron – 4.8; 0.9 m; depth, water apron teeth width – 1.19; 0.40 m; length, thickness of the rear apron – 20.36; 0.6 m; length of the tubular part of the regulator – 16.3 m) taking into account the placement of gates, service and road bridges. The shear stability of the floor and the structure as a whole was determined.
Recommender systems are a valuable tool for software engineers. For example, they can provide developers with a ranked list of files likely to contain a bug, or multiple auto-complete suggestions for a given method stub. However, the way these recommender systems interact with developers is often rudimentary -- a long list of recommendations only ranked by the model's confidence. In this vision paper, we lay out our research agenda for re-imagining how recommender systems for software engineering communicate their insights to developers. When issuing recommendations, our aim is to recommend diverse rather than redundant solutions and present them in ways that highlight their differences. We also want to allow for seamless and interactive navigation of suggestions while striving for holistic end-to-end evaluations. By doing so, we believe that recommender systems can play an even more important role in helping developers write better software.
Sherlock A. Licorish, Christoph Treude, John Grundy
et al.
Six months ago an important call was made for researchers globally to provide insights into the way Software Engineering is done in their region. Heeding this call we hereby outline the position Software Engineering in Australasia (New Zealand and Australia). This article first considers the software development methods practices and tools that are popular in the Australasian software engineering community. We then briefly review the particular strengths of software engineering researchers in Australasia. Finally we make an open call for collaborators by reflecting on our current position and identifying future opportunities
The rapid advances in technology over the last decade have significantly altered the nature of engineering knowledge and skills required in the modern industries. In response to the changing professional requirements, engineering institutions have updated their curriculum and pedagogical practices. However, most of the changes in the curriculum have been focused on the core engineering courses without much consideration for the auxiliary courses in mathematics and sciences. In this paper, we aim to propose a new, augmented mathematics curriculum aimed at meeting the requirements of the modern, technology-based engineering workplace. The proposed updates require minimal resources and can be seamlessly integrated into the existing curriculum.
Abstract Recently, the application and development of flexible microwave-absorption composites based on silicone rubber have gradually become a research hot spot. In this study, methyl vinyl phenyl silicone rubber (MPVQ)/carbonyl iron particles (CIPs)/graphene (GR) composites were prepared by mechanical blending, and the effects of thermal-ageing temperature on the microwave-absorption properties of the composites were investigated. The mechanism of the thermal-ageing temperature’s effects on microwave-absorption behaviour was identified. The results show that unaged composites have superior microwave-absorption properties, with a minimum reflection loss (RL min ) of − 87.73 dB, a lowest thickness of 1.46 mm, and an effective absorption bandwidth (EAB, RL < − 10 dB) reaching 5.8 GHz (9.9–15.7 GHz). With ageing at 240 °C for 24 h, the RL min at a frequency of 5.48 GHz is − 45.55 dB with a thickness of 2.55 mm, and the EAB value reaches 2 GHz (range 4.6–6.6 GHz). In the thermal-ageing process, a crosslinking reaction occurs in MPVQ with an increase in crosslinking density from 5.88 × 10−5 mol g−1 (unaged) to 4.69 × 10−4 mol g−1 (aged at 240 °C). Simultaneously, thermal degradation of the composites leads to a reduction in the rubber concentration. In addition, a small amount of CIPs are oxidized to Fe3O4, and the remaining CIPs aggregate to generate more electrically conductive pathways. Consequently, the dielectric loss of the composites will be significantly improved, resulting in poor impedance matching. The microwave-absorption properties of the composites gradually decrease with increasing thermal-ageing temperature from 200 to 240 °C.