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.
Objective Current ship navigation decision-making systems struggle to demonstrate superior performance in undefined sailing scenarios. Given the broad applicability of large language models （LLMs） in unknown scenarios, this study proposes a dual-LLM-core-driven adaptive ship navigation agent architecture （Nav-DLLC） to address this issue.MethodNav-DLLC employs ReAct-based prompting to decompose complex navigation tasks into manageable subtasks and invoke external tools for information collection, reducing LLM errors. Subsequently, a small-parameter LLM fine-tuned with low-rank adaptation （LoRA） serves as the collision avoidance core, processing unstructured data to generate COLREG-compliant decisions. ResultsSimulation experiments show that Nav-DLLC achieves outstanding performance in both traditional ship collision avoidance tasks and unstructured dynamic scenarios. Its collision avoidance accuracy is 86%, and its behavior compliance rate is 90%, significantly outperforming LLM baselines and traditional methods such as the Dynamic Window Approach （DWA） and Artificial Potential Field （APF）. The decision core's single-decision latency is 11.13 seconds, higher than the 0.73 seconds of traditional methods, yet still within the safe time window for collision avoidance decision-making. ConclusionNav-DLLC bridges the gap between traditional navigation systems and LLM technology, providing a safe and efficient intelligent decision-making paradigm for complex navigation environments and promoting the intelligent development of ship navigation.
This study examines the potential distribution of Mackerel scad (<i>Decapterus macarellus</i>) in the South China Sea under future climate scenarios (SSP 1.26, SSP 2.45, SSP 5.85) using an ensemble species distribution model (SDM). Key environmental variables included sea surface salinity (SSS), sea surface height (SSH), sea surface temperature (SST), mixed-layer depth (MLD), chlorophyll-a concentration (CHL), and sea-bottom temperature (SBT). Results show that SST and MLD are the primary drivers of habitat suitability, with current suitable habitats concentrated in the northern offshore areas. Projections for the 2050s and 2090s indicate a reduction in suitable habitats, particularly under high-emission scenarios, with more gradual reductions under low-emission scenarios. Habitat loss is most pronounced in the northern South China Sea, while the central region is projected to see an expansion of suitable habitats. These findings highlight the climate impact on <i>D. macarellus</i> distribution and inform sustainable management strategies for the species in the region.
With the growing deployment of multi-USV (unmanned surface vehicle) systems for complex maritime operations, coordinated path planning is critical for safety and efficiency in congested waterways. Classical multi-agent path finding (MAPF) methods, however, often neglect vessel kinematics and collision envelopes, yielding trajectories that are impractical or unsafe at sea. We present a centralized planning framework based on any-angle Safe Interval Path Planning (AA-SIPP) augmented with a vessel-specific maximum yaw-rate constraint. This yields smooth, kinematically feasible trajectories while preserving continuous-time separation. The approach is validated in high-fidelity Gazebo marina simulations involving up to 20 USVs based on the WAM-V platform. Compared with Conflict-Based Search (CBS), a representative grid-based MAPF algorithm, our framework maintained the prescribed safety distances and achieved zero collisions across all scenarios considered, whereas CBS exhibited separation violations in simulation. The method also scales well: mission makespan remained nearly constant as fleet size increased. These results support the applicability of dynamically constrained MAPF to maritime coordination in congested environments.
Recently, the exponential growth in capability and pervasiveness of Large Language Models (LLMs) has led to significant work done in the field of code generation. However, this generation has been limited to code snippets. Going one step further, our desideratum is to automatically generate architectural components. This would not only speed up development time, but would also enable us to eventually completely skip the development phase, moving directly from design decisions to deployment. To this end, we conduct an exploratory study on the capability of LLMs to generate architectural components for Functions as a Service (FaaS), commonly known as serverless functions. The small size of their architectural components make this architectural style amenable for generation using current LLMs compared to other styles like monoliths and microservices. We perform the study by systematically selecting open source serverless repositories, masking a serverless function and utilizing state of the art LLMs provided with varying levels of context information about the overall system to generate the masked function. We evaluate correctness through existing tests present in the repositories and use metrics from the Software Engineering (SE) and Natural Language Processing (NLP) domains to evaluate code quality and the degree of similarity between human and LLM generated code respectively. Along with our findings, we also present a discussion on the path forward for using GenAI in architectural component generation.
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.
The violent swinging motion of a payload imposes significant safety problems for the operation of offshore ship-mounted cranes (offshore cranes). We therefore propose a stable experimental hoisting platform for an offshore crane payload positioning system with a parallel cable-driven method (PP-PCDM), and an adaptive cable-drive anti-swing tension (ACAT) control method based on the PP-PCDM is developed to resolve the problem of swinging of the payload by limiting its spatial position. The PP-PCDM enables synchronous tracking of the movement of the payload when the crane is working. When the payload deviates from a stable state due to an external disturbance, the anti-swing cables are immediately retracted or released based on the feedback signal. The spatial position of the payload is then limited by adjusting the length and tension of the parallel cables until the payload appears stationary from the perspective of the ship’s deck. Operational safety and efficiency is substantially improved, and the proposed PP-PCDM structure and ACAT control method can be applied to a variety of different types of cranes. The results of simulations and physical experiments show that the anti-swing effect exceeds 89.86%. The PP-PCDM enables excellent performance of synchronous tracking and hoisting assistance, and ensures that the rated lifting weight of the offshore crane is not affected.
<i>Pyropia yezoensis</i> is the most widely cultivated and economically important alga. Affected by climate change, the cultivation of <i>P. yezoensis</i> has gradually migrated to the northern coast of China, increasing from 6.8% in 2019 to 19.5% in 2023. To date, the genetic impact of northern migration on cultivation resources has not been assessed and analyzed extensively. Here, DNA barcoding (<i>rbc</i>L and <i>cox</i>1) and the presence/absence of intronic-ORFs in mitochondrial regions (<i>rnl</i> and <i>cox</i>1) were applied to investigate genetic diversity in 44 <i>P. yezoensis</i> specimens from 17 aquaculture farms in China, with comparisons to Korean and Japanese cultivated resources. The lower intraspecific variation was 0.31% for the <i>cox</i>1 gene and 0.14% for the <i>rbc</i>L gene, with three haplotypes, indicating that intensive selection and breeding during cultivation had narrowed the germplasm genetic variation. The intron structure of mitochondrial regions showed that the cultivated resources had 17 phenotypes, and the northern specimens shared 35.3% of genotypes with the southern specimens, indicating that the cultivated <i>P. yezoensis</i> is expanding its cultivation ranges through north migration. Even with lower genetic diversity, the northern area of cultivation had already developed 17.6% site-specific specimens. The genetic diversity of cultivated <i>P. yezoensis</i> from the Northwest Pacific is also discussed. Our work provides a preliminary framework for <i>P. yezoensis</i> breeding and cultivation under climate change.
This study introduces a groundbreaking approach for real-time 3D localization, specifically focusing on achieving seamless and precise localization during the terminal guidance phase of an autonomous underwater vehicle (AUV) as it approaches an omnidirectional docking component in an automated deployable launch and recovery system (LARS). Using the AUV’s magnetometer, an economical electromagnetic beacon embedded in the docking component, and an advanced signal processing algorithm, this novel approach ensures the accurate localization of the docking component in three dimensions without the need for direct line-of-sight contact. The method’s real-time capabilities were rigorously evaluated via simulations, prototype experiments in a controlled lab setting, and extensive full-scale pool experiments. These assessments consistently demonstrated an exceptional average positioning accuracy of under 3 cm, marking a significant advancement in AUV guidance systems.