Abstract
The automated detection of deep-sea polymetallic nodules is critical for processing large volumes of benthic imagery. However, its scalability faces challenges from cross-expedition covariate shifts, such as changes in lighting, altitude, and camera payloads, which lower zero-shot model performance. While semi-supervised pseudo-labelling presents a potential alternative to time-consuming re-annotation, simple implementations can quickly lead to confirmation bias. This study identifies two primary sources of this degradation: spatial noise from tiling fragmentation at tile borders and an architecture-agnostic interior false positive floor caused by semantic domain shift. This work proposes using a multi-model ensemble for pseudo-labelling to reduce the noise impact. Using a spatial border filter and confidence stratification, three architecturally distinct teacher models (YOLOv8, Faster R-CNN, and DINO) are employed to determine a reliable and domain-invariant subspace. Under a strict anti-leakage Leave-One-Partition-Out protocol, the proposed approach surpasses the supervised fine-tuning baseline at 100-tile pseudo-label budget across four random seeds (macro mAP50:95 of 0.4745 +/- 0.0042 versus 0.4467 +/- 0.0079), with gains concentrated in the most domain-shifted fold. Beyond this budget, our findings highlight two important adaptation trends: a pool-size degradation trend where excessive pseudo-label volume actively degrades generalisation, and the observation that the fine-tuned models reduce pseudo-label fidelity despite higher precision, providing evidence for the advantage of using frozen source checkpoints for cross-domain adaptation.

Abstract
Understanding how structural change drives long-run growth requires jointly considering the dynamics of productive and scientific specialisations, and science-industry alignment. This paper develops and tests a unified framework that integrates evolutionary, structuralist, complexity, and innovation-systems perspectives to assess how productive and scientific specialisations, science-industry alignment, diversification, and global value chain integration shape economic performance. To operationalize this framework, we construct new indicators, including a Science-Industry Matching (SIM) index, measures of dynamic entry and relatedness density, and specialisation-based diversity indices, and apply them to a panel of up to 142 countries over 2000-2018/2023. Estimation relies on country fixed effects with Driscoll-Kraay standard errors to address heteroskedasticity, autocorrelation, and cross-sectional dependence. The results reveal that persistent specialisation in high- and medium-high-tech industries fosters growth, while low-tech dependence constrains it. Scientific specialisation in enabling fields such as mathematics, physics, chemistry, and energy/environmental sciences supports growth, but excessive concentration risks lock-in. Science-industry alignment enhances growth in advanced economies with strong absorptive capacity but penalises weaker systems. Industrial diversification often dilutes resources, whereas scientific diversification consistently promotes growth by broadening the knowledge base for recombination. Finally, integration into global value chains is growth-enhancing in developing economies, while advanced economies can sustain higher domestic value added without significant penalties.

Abstract
In recent years, the research community, but also the general public, has raised serious questions about the reproducibility and replicability of scientific work. Since many studies include some kind of computational work, these issues are also a technological challenge, not only in computer science, but also in most research domains. Computational replicability and reproducibility are not easy to achieve due to the variety of computational environments that can be used. Indeed, it is challenging to recreate the same environment via the same frameworks, code, programming languages, dependencies, and so on. We propose a framework, known as SciRep, that supports the configuration, execution, and packaging of computational experiments by defining their code, data, programming languages, dependencies, databases, and commands to be executed. After the initial configuration, the experiments can be executed any number of times, always producing exactly the same results. Our approach allows the creation of a reproducibility package for experiments from multiple scientific fields, from medicine to computer science, which can be re-executed on any computer. The produced package acts as a capsule, holding absolutely everything necessary to re-execute the experiment. To evaluate our framework, we compare it with three state-of-the-art tools and use it to reproduce 18 experiments extracted from published scientific articles. With our approach, we were able to execute 16 (89%) of those experiments, while the others reached only 61%, thus showing that our approach is effective. Moreover, all the experiments that were executed produced the results presented in the original publication. Thus, SciRep was able to reproduce 100% of the experiments it could run.

Abstract
Self-service technologies (SSTs) that replace regular checkout are widely deployed in grocery retailing to reduce customer frictions and labor needs, yet their impact on consumer purchasing behavior remains unclear. We study a specific type of these SSTs: mobile app 'scan & go' technologies. Mobile app scan & go technologies may lower customer time and effort spent during a shopping trip by eliminating queuing and double handling. However, they also shift scanning effort to customers, which may change attention and shopping patterns. We explore how customers adopt mobile app scan & go technologies in practice, and study the causal effect of adoption on their purchasing behavior. We partner with a European grocery retailer that introduced a mobile app scan & go technology in their physical stores and analyze a large transactional dataset spanning more than 7 million purchases from nearly 60,000 customers. Leveraging the staggered adoption timing and matched non-adopters, we estimate the effect of adoption using difference-in-differences designs with customer and time fixed effects and modern staggered-DiD estimators. We find that adoption increases customers' monthly purchase frequency and total monthly spending, with little change in average basket value. We also find that adoption is not 'all-or-nothing': most adopters use the scan & go technology selectively for particular shopping trips, and about one-third try it once and then discontinue using it. The observed spending and frequency gains are concentrated among customers who use the technology repeatedly. These findings suggest that mobile app scan & go technologies can strengthen customer retention, but only when they reliably reduce customer friction. Retailers should promote mobile app scan & go usage for 'major' grocery trips and design onboarding and in-store support to reduce first-use learning costs - because repeated usage, not mere adoption, is what drives performance benefits.

Abstract

Abstract
Researchers rely on confidential and sensitive microdata provided by national statistical institutes and other organizations, making disclosure control a critical challenge. Manual output checking processes are time-consuming and require expert knowledge, limiting scalability. This paper presents an automated framework that integrates large language models (LLMs), prompt engineering, and workflow automation (n8n) for statistical disclosure control (SDC). The system introduces an AI-driven output validation, including code generation, data processing, and risk assessment, allowing researchers to pre-check their outputs via a Seamless Within-Activity Review (SWAR) approach. Key challenges such as computational costs, confidentiality concerns, and the need for human oversight are addressed, and reinforcement learning is proposed to enhance future risk evaluation. The framework marks a step toward scalable, privacy-compliant, AI-assisted disclosure control in official statistics.