Search Results (2,426)

People with special needs often face barriers to participating in adapted outdoor leisure physical activities. A participatory action research project involving a nonprofit organization, a citizen with motor disabilities, and researchers aimed to co-develop a digital platform connecting people with special needs interested in outdoor leisure physical activities with trained volunteers. The adopted co-design methodology followed four stages: (1) Exploration (identifying users’ needs and preferences), (2) Co-design (defining key information and platform features), (3) Validation (prioritizing features), and (4) Development (implementing and testing the platform). This article focuses on stages 2, 3, and 4. During stage 2, key information and features were identified to support matching people with special needs and volunteers and informing users about adapted outdoor leisure physical activities. In stage 3, these elements were prioritized using eight key considerations, including technological (e.g., ease of use), environmental (e.g., avoiding redundancy with existing initiatives), organizational (e.g., availability of human resources), and financial factors (e.g., grant planning). Stage 4 resulted in the launch of Tandem Actif, followed by user testing to document user experience and guide improvements. This article details the application of co-design within a participatory action research project aimed at promoting safe, ethical, and accessible participation in outdoor leisure physical activities for people with special needs. Full article

This systematic review evaluates recent scientific and technological advances in water quality monitoring and pollution alarms for ports, based on records retrieved from seven databases following the PRISMA protocol. A total of 414 documents were screened, resulting in 141 articles (TRL 3) and 56 patents (TRL 4–5). Bibliometric, patentometric, and thematic analyses were conducted using Bibliometrix and ORBIT^®. Results show sustained growth in both academic and technological outputs, with a patent Compound Annual Growth Rate (CAGR) of 32%, compared with 13% for scientific publications, indicating accelerated translation from research to innovation. The conversion rate from scientific research to patenting increased from 14% (2010–2015) to 47% (2020–2023). Analysis of patent legal status reveals that 52% of patent families remain valid (48% granted; 4% pending), while 33% are lapsed, 13% revoked, and 2% expired, reflecting the dynamic and emerging character of the field. Technological ownership is highly concentrated, with China accounting for nearly all active patents, whereas scientific production is more geographically distributed. Thematic analysis identifies four main scientific clusters: environmental monitoring, chemical pollutants, seashore hazards, and eutrophication. The main technological domains of the patents are analysis of biological materials, control, and environmental technologies. Emerging areas of focus at TRL 3 and TRL 4–5 include microplastics, climate-change impacts, aquaculture risks, real-time sensing, IoT-enabled platforms, machine-learning analytics, autonomous monitoring systems, and bioindicator-based early-warning tools. This review provides a quantitative roadmap to support sustainable port operations, coastal ecosystem protection, and progress toward multiple synergistic United Nations Sustainable Development Goals (SDGs). Full article

Vehicle-to-everything communication grants connected and automated road vehicles the opportunity to share their sensor information such as detected road objects for collective awareness. This paper compares various state fusion strategies within a high-level cooperative perception architecture, focusing on the fusion of object-level information provided in standard Collective Perception Messages. This work compares five track-to-track fusion methods, namely Covariance Intersection, Inverse Covariance Intersection, Adapted Extended Kalman Filter, Adapted Unscented Kalman Filter and Information Matrix Fusion, using a simulation framework built with CARLA and Autoware. The methods are analyzed in a case study to assess their performance under different vehicle maneuvers and varying input information accuracy. The case study highlights trade-offs between fusion strategies and illustrate their behavior in asynchronous multi-agent scenarios. While the analysis is conducted in simulation, the architecture is designed to be extensible, and directions for future development are outlined, including the integration of classification and object confidence fusion modules. Full article

Machine Learning (ML) models trained on large-scale datasets learn useful predictive patterns, but they may also memorize undesired information, leading to risks such as information leakage, bias, copyright violations, and privacy attacks. As these models are increasingly deployed in real-world and regulated settings, the consequences of such memorization become practical and high-stakes, reinforced by data-protection frameworks that grant individuals a Right to be Forgotten (e.g., the GDPR). Simply removing a record from the training dataset does not guarantee the elimination of its influence from the model, while retrain-from-scratch procedures are often prohibitive for modern architectures, including Transformers and Large Language Models (LLMs). In this work, we provide a perspective on Machine Unlearning (MU) in supervised learning settings, with a particular focus on Natural Language Processing (NLP) scenarios, grounded in a PRISMA-driven systematic review. We propose a multi-level taxonomy that organizes MU techniques along practical and conceptual dimensions, including exactness (exact versus approximate), unlearning granularity, guarantees, and application constraints. To complement this perspective, we run an illustrative benchmark evaluation using a standardized unlearning protocol on DistilBERT trained on a public corpus of news headlines for topic classification, contrasting the retraining gold standard with representative design-for-unlearning and approximate post hoc techniques. For completeness, we also report two oracle-assisted upper-bound baselines (distillation and scrubbing) that rely on a clean retrained reference model, and we account for their incremental cost separately. Our analysis jointly considers model utility, probabilistic quality, forgetting and privacy indicators, as well as computational efficiency. The results highlight systematic trade-offs between accuracy, computational cost, and removal effectiveness, providing practical guidance for selecting machine unlearning techniques in realistic deployment scenarios. Full article

Adaptive reuse has emerged to become a tool for implementing the understanding of sustainability in the domain of architectural conservation, as it encourages the continued usage of old buildings as means of reducing environmental impact, as well as preserving socio-cultural capital while generating economic income. However, in its practice, the decisions regarding granting meanings, interpretation, and preserving memories within adaptation processes are dominated by expert-driven approaches that inadequately incorporate stakeholder values or intangible heritage dimensions. To this end, this study aims to contribute to the current debate by adopting a participatory co-evaluation framework that integrates both authenticity perspectives and sustainability dimensions using Multi-Criteria Decision Analysis (MCDA) for evaluating adaptive reuse alternatives for an abandoned prefabricated wooden heritage building. Stakeholder priorities were drawn through a workshop and transformed into normalized weights using the Simos technique. Four design alternative typologies—namely, Continuity, Cultivation, Differential, and Optimization—were assessed and compared against 20 performance indicators across heritage, social, ecological, and economic criteria using the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). Indicator-level analyses and sensitivity tests (±10% and ±20% weight variations) were applied to confirm the robustness of rankings. The results from the best-performing alternative demonstrated the trade-offs between heritage authenticity and sustainability objectives, as well as demonstrating how combining participatory methods with quantitative evaluation can support evidence-based decision-making for adaptive reuse. The applied integrated framework helps bridge the gap between heritage theory and practice by combining authenticity, participation, and sustainability in one analytical approach, supporting evidence-based decisions for adaptive reuse. Full article

Cartilage is an important yet vulnerable tissue with limited self-healing capacity, where damage often progresses to joint degeneration, which eventually leads to severe osteoarthritis (OA). Current tissue engineering strategies focus on biocompatible scaffolds for cartilage regeneration, particularly amnion (or amniotic membrane), emerging as a promising biomaterial due to its wide availability, low immunogenicity, and naturally derived microenvironment that is advantageous for cartilage regeneration. This systematic review aims to evaluate the existing evidence on the efficacy of amnion as a tissue scaffolding material for cartilage regeneration in both preclinical and clinical studies. Using terms such as “cartilage damage”, “cartilage injuries”, “amnion” and “amniotic membrane”, 19 relevant studies were identified across three major databases (PubMed, Scopus and Web of Science) until 25 December 2025. All preclinical and clinical studies that utilized amnion for cartilage repair or as cartilage tissue engineering scaffolding materials were included. Evidence quality was assessed using the OHAT and MINORS risk of bias tool. This study is prospectively registered in the PROSPERO database under the ID 1178444. The findings consistently indicate that amniotic scaffolds, regardless of processing methods or cell seeding, yield favorable outcomes without adverse effects across different species. In vitro analysis revealed that treatment groups with amnion show better cell attachment, viability, and proliferation, and higher content of cartilage-related markers expressed by the seeded cells, either chondrocyte, bone marrow-derived mesenchymal stem cells (MSCs), adipose tissue-derived MSCs, placenta-derived MSCs, umbilical cord-derived MSCs, amniotic MSCs or amniotic epithelial cells. In in vivo and ex vivo studies, amnion-treated groups demonstrated improved quality of the treated cartilage, with better integration, as indicated by higher histological scores and the presence of type II collagen (COL-II). There was an inconsistency in the reporting of cartilage defect dimensions in the in vivo models across the different studies. Nevertheless, the outcome measurements were consistently reported with histological analysis, with or without International Cartilage Repair Society (ICRS) scoring and immunohistochemistry (IHC) analysis, across the studies. Clinically, most subjects show improvement in the Knee Injury and Osteoarthritis Outcome Score (KOOS) Sports and Recreation score and KOOS Quality of Life score, as well as reduced Visual Analogue Scale (VAS) average and maximum pain scores. In conclusion, preclinical and clinical studies support amnion as an ideal scaffold material for cartilage tissue engineering and regeneration. Future research should focus on optimizing and standardizing amnion scaffold preparation at a production scale to facilitate the translation of these positive outcomes into clinical applications. This study is funded by the Ministry of Higher Education Malaysia via Prototype Research Grant Scheme (PRGS/1/2021/SKK01/UM/02/1) and UM International Collaboration Grant—2023 SATU Joint Research Scheme Program: ST007-2024. Full article

Climate change significantly affects hydrological responses, yet studies addressing future water availability in the Paraopeba River Basin (PRB), an important tributary of the São Francisco River Basin in Brazil, remain limited, particularly under CMIP6 scenarios and using distributed hydrological modeling approaches. In this context, this study evaluated the hydrological responses of the PRB, under climate change using the MHD-INPE. Future projections were based on an ensemble of seven climate models from the NEX-GDDP-CMIP6 collection, considering a baseline period (1992–2014), three future periods 17(2040–2060, 2061–2080 and 2081–2100) and two socioeconomic scenarios (SSP245 and SSP585). The model satisfactorily reproduced observed streamflow during the baseline period. Under the SSP585 scenario, the projections indicate stronger alterations in water availability, with a potential intensification of flood and drought events, as reflected by reductions in minimum streamflows (Q₉₀) and increases in maximum streamflows (Q₁₀), particularly in sub-basins 4 and 5, where Q₉₀ reductions approach 30% and Q₁₀ increases reach 11.7%. Additionally, a decrease in Q_7,10 values was observed, which enabled the analysis of the Conflict Index (Icg), indicating that water withdrawals currently granted may exceed the limits established by existing legislation in future scenarios (Igc > 1). Full article

The consolidation of artificial intelligence (AI) is transforming the journalistic sector, to the point that its ethical dimension is being altered. However, the mission and values of the media in the face of the current emergence of generative artificial intelligence (GenAI) have barely been explored. Bearing this in mind, it is important to understand not only how journalists perceive AI, but also to examine the role that the media assign to themselves and the audience’s participation in this context. This research explores the roles defined by a sample of leading media outlets (n = 21) in seven countries in Western Europe and North America: France, Germany, Italy, Spain, the United Kingdom, Canada, and the United States. To this end, a discursive content analysis is applied to three newspapers (printed or digital) per country. The findings reflect differences between countries and media outlets, within a common trend of prioritizing responsibility as the primary editorial value, followed by truthfulness. We also found scant direct references to AI regulation, alongside the development of participatory interactivity within readership established by the media outlet. Furthermore, greater participation of audiences was observed in publicly funded publications, granting audiences a deliberative role. Full article

Background: Food insecurity is defined as having limited access to food, while nutrition insecurity is characterized as a lack of consistent access to affordable and acceptable foods that support health, manage or prevent disease, and meet daily nutritional needs. College students face increased risks of food and nutrition insecurity, yet the issue is understudied. This study examined the patterns of food and nutrition insecurity among students at a public university across main and regional campuses. Methods: We conducted a cross-sectional survey using Qualtrics for participant recruitment in November 2023. The USDA’s 10-item toolFood insecurity was measured using the USDA’s 10-item tooland housing security was measured using the U.S. Census Bureau’s National Survey of Income and Program Participation 6-item tool. We used ANOVAs and logistic regression to examine differences across demographics. Data analysis was done using SPSS version 29. Results: There were 6538 student responses. Of these, 36% of students were food insecure, while 20% were nutritionally insecure. Comparatively, food and nutrition insecurity were significantly higher among students with low-income (p < 0.001), housing instability (p < 0.001), a higher number of dependents, and those indicating that they were single/unmarried (p = 0.005), first-generation (p < 0.001), and Pell grant eligible (p < 0.001). Annual income and housing security emerged as significant predictors: lower income was approximately twice as likely to be associated with food insecurity, while those reporting housing insecurity were six times more likely to experience food insecurity. Conclusions: The study findings reveal disparities in food and nutrition insecurity among a diverse student population at a public university. Addressing the issue among them is crucial and requires a multifaceted, inclusive approach. Emergency financial assistance and structural interventions that promote housing security are warranted. Full article

Within the project “Small-scale grants for biodiversity actors in South-East Europe 2023–2025”, whose main mission is the collection of local varieties and populations, a fundamental question arises: “Why are ‘ancestral’ maize varieties and populations still cultivated?” To answer this question, we conducted a comprehensive set of investigations on 14 maize (Zea mays L.) varieties and populations and on one hybrid, collected from the historical regions of Romania—Transylvania, Moldova, Oltenia and Muntenia. The studies combine quantitative, qualitative and computational methods and focused on energy consumption associated with maize cultivation; maize grain production and related agronomic characteristics; the content of energy macronutrients (protein, fat, starch) and energy elements (CHNS-O, ash, moisture); and nutritional and thermal energy values (upper and lower) of whole cornmeal. The sustainability of the cultivation of local maize varieties was also evaluated based on the energy balance and the energy efficiency ratio. The results demonstrated that the cultivation of “ancestral” maize varieties and populations is sustainable, because the amount of energy obtained, expressed in kJ ha⁻¹, as nutritional energy (24,740,195.04–90,287,743.07), higher heating energy (55,162,983.798–193,374,572.55) and lower heating energy (32,329,465.37–113,906,753.63), is greater than the amount of energy consumed for the establishment and maintenance of these crops (1,742,798.75–19,524,555.05). Full article

As artificial intelligence becomes increasingly embedded in hospitality and tourism services, it is reshaping employees’ innovative work behavior. Grounded in the Job Demands-Resources perspective, this study examines how AI self-efficacy affects innovative work behavior and proposes a moderated mediation model to investigate the mediating role of work engagement and the boundary condition of schedule idiosyncratic deals. Using a three-wave time-lagged design, the study collected data from 300 employees working in the hospitality and tourism industry in Korea. The findings show that AI self-efficacy positively predicts innovative work behavior both directly and indirectly through increased work engagement. Furthermore, this mediating process is strengthened by higher levels of schedule i-deals, confirming a positive moderating effect. Theoretically, this study extends human-AI collaboration research by broadening the explanatory scope of the Job Demands-Resources model in the AI context. Practically, organizations undergoing digital transformation should provide training that strengthens employees’ confidence in using AI and grant greater autonomy over work schedules. Such practices help create a supportive environment that enables AI self-efficacy to translate into work engagement and ultimately innovative work behavior. Full article

This paper presents qualitative findings from a broader study conducted with beneficiaries of social assistance grants in Nkonkobe Municipality, South Africa, guided by Amartya Sen’s Capability Approach and a Feminist Political Economy perspective. It specifically examines the impact of the Child Support Grant (CSG) on women residing in rural areas with limited economic opportunities. Drawing on the narratives of twenty-five female beneficiaries, the study explores the grant’s role beyond child welfare outcomes. Despite its small monetary value, the CSG demonstrated significant and unintended transformative effects. Within contexts of persistent poverty, unemployment, and inequality, the grant contributed to alleviating household hardship and enhancing women’s empowerment. Participants reported increased economic agency, autonomy, and decision-making capacity. The CSG also served as a critical resource for women seeking to exit abusive relationships by providing a measure of financial independence, often lacking among survivors of domestic violence. These findings contribute to broader debates on welfare services as tools for equality, diversity, and democracy, highlighting the CSG’s potential to advance gender empowerment and social inclusion. However, the grant alone remains insufficient to ensure sustainable economic security. The paper recommends that the CSG be strengthened through increased value and complemented by active labour market interventions that promote women’s economic participation. Full article

Predicting the future impact of technical economic documents at release time is challenging due to delayed supervision signals, long-tailed label distributions, and time- and domain-dependent shifts in language and topics. Moreover, similarity graphs derived from text embeddings can be noisy due to boilerplate and evolve under temporal drift, making robustness and leakage-free evaluation essential. We formulate grant-time patent impact prediction as a node classification and within-domain ranking problem on a large-scale semantic similarity document graph built from patent text embeddings, avoiding any future citation leakage. The document graph is constructed via ANN Top-K retrieval and similarity thresholding, enabling scalable and reproducible sparsification on hundreds of thousands of nodes. We propose GraphGPT-Patent, which adapts a reversible graph-to-sequence foundation backbone to local subgraphs extracted from the similarity network. The model incorporates time- and domain-conditioned edge reliability to suppress drift-induced and template-driven pseudo-similarity, and optimizes a joint objective coupling high-impact classification with ranking consistency within comparable groups. Experiments on USPTO granted patents (2000–2022) across three high-volume CPC domains and three evaluation horizons show consistent gains over text-only and GNN baselines, achieving up to 0.94 recall for the positive class and improved macro-average recall across nine settings. Temporal shift analyses further quantify the effect of training-data freshness, while explanation subgraphs provide auditable structural evidence of model decisions. The proposed framework offers an effective graph-based learning pipeline for scalable impact prediction and downstream triage under strict information constraints. Full article

Background: Fixed-dose combination drugs (FDCs) are combinations of two or more active ingredients in a single dosage form. These formulations have proven effective in combating the development of resistance in diseases such as tuberculosis and malaria. Despite the benefits observed in the aforementioned cases, fixed-dose antibiotics combinations (FDACs) are increasingly raising questions about their rationality. This is the case for several FDACs listed in the AWaRe classification as not recommended, which unfortunately remain available on the pharmaceutical market, particularly in low- and middle-income countries like the Democratic Republic of Congo (DRC). Objectives: To identify the essential medicines available in pharmacies open to the public in the city of Kinshasa and to assess their inclusion in the DRC’s National List of Essential Medicines (NLEM) and in the World Health Organization’s (WHO) List of Essential Medicines (LEM). The rationality of the FDACs circulating in the city of Kinshasa were also evaluated based on the 2023 AWaRe classification. Methods: A cross-sectional and descriptive study was conducted between February and October 2025 in Kinshasa. For this purpose, fifty registered pharmacies open to the public were selected by systematic random sampling as the research sample. Data collection consisted of completing a data collection form after we had provided the pharmacies’ owners with the necessary explanations regarding the importance of the study and guaranteed their anonymity. Results: The controlled FDACs encountered comprised 27 specialties across 15 different formulations. Out of 15 formulations, 12 (80%) were included on the WHO list of non-recommended antibiotics and were not included in the DRC’s NLEM nor in the WHO’s LEM. Some had been withdrawn from the market in their countries of manufacture. Of the 15 FDACs evaluated for their rationality and compliance, the injectable FDACs presented problems related to the relevance and completeness of information contained on their packaging. On their primary packaging, there was a significant difference in the expiration dates of the powder and sterile water for injection contained in the combination pack, ranging from 6 to 36 months. Furthermore, the secondary packaging lacked data related to the sterile water for injection contained in the combination pack. In addition, several medications contained the same therapeutic combination. For injectable FDAC, for example, the combination Ceftriaxone-Sulbactam was represented by eight medications. For oral FDACs, the combination Sulfamethoxazole-Trimethoprim was represented by seven medications. Globally, 100% of these drug combinations originated from India. Conclusions: Fifteen varieties of FDACs were available in Kinshasa, most of which (80%) were unsuitable. It is important that public health authorities address this situation and develop stricter guidelines for granting marketing authorizations, particularly for FDACs. Full article

Ovarian cancer is a devastating disease that is often diagnosed in the late stages. The typical therapeutic approach includes surgery plus cytotoxic drugs such as carboplatin and paclitaxel. In recent years, the advent of poly ADP-ribose polymerase (PARP) inhibitors such as olaparib has offered additional treatment opportunities for patients with BRCA mutations or homologous recombination deficiencies. Nevertheless, resistance to therapy usually occurs, leading to poor overall survival. Therefore, novel treatments are needed for this disease. One of the obstacles to successful treatment is the highly immunosuppressive nature of the ovarian cancer microenvironment. Recent strategies for the treatment of ovarian cancer and other types of cancer involve targeting the metabolism of cancer cells and other cells of the tumor microenvironment. One drug that has been investigated both in preclinical studies and clinical trials as an antitumor agent is metformin. This drug, typically used for the treatment of type-2 diabetes for its capability to lower blood glucose, can directly affect cancer cell growth and survival by activating the AMPK (adenosine monophosphate-activated protein kinase) pathway. Furthermore, it can affect the phenotype of other cells of the tumor microenvironment such as macrophages and T cells. In this review, we summarize the main characteristics of ovarian cancer and describe preclinical studies and clinical trials involving metformin as a therapeutic agent for this disease. Full article