Data-driven neural word embeddings (NWEs), grounded in distributional semantics, can capture various ranges of linguistic regularities, which can be further enriched by incorporating structured knowledge resources. This work proposes a novel post-processing approach for injecting semantic relationships into the vector space of both static and contextualized NWEs. Current solutions to retrofitting (RF) word embeddings often oversimplify the integration of semantic knowledge, neglecting the nuanced differences between relationships, which may result in suboptimal performance. Instead of applying multi-thresholding to distance boundaries in metric learning, we compute taxonomic similarity to dynamically adjust these boundaries during the semantic specialization of word embeddings. Benchmark evaluations on both static and contextualized word embeddings demonstrate that our dynamic-fitting (DF) approach produces SOTA correlation results of 0.78 and 0.76 on SimLex-999 and SimVerb-3500, respectively, highlighting the effectiveness of incorporating multiple semantic relationships in refining vector semantics. Our approach also outperforms existing RF methods in both supervised and unsupervised semantic relationships recognition tasks. It achieves top accuracy scores for hypernymy detection on the BLESS, WBLESS, and BIBLESS datasets (0.97, 0.89, and 0.83, respectively) and an F1 score of over 0.60 on four types of semantic relationship classification in the shared Subtask-2 of CogALex-V, surpassing all participant systems. In the analogy reasoning task of the Bigger Analogy Test Set, our approach outperforms existing RF methods on inferring relational similarity. These consistent improvements across various lexical semantics tasks suggest that our DF approach can effectively integrate distributional semantics with symbolic knowledge resources, thereby enhancing the representation capacity of word embeddings in downstream applications.

Despite worldwide uptake, there has been little published evaluation of actually delivering the World Health Organization (WHO) Mental Health Gap Action Programme (mhGAP) in typical low- and middle-income countries (LMICs). This paper aims to evaluate the impact of a pilot study in which mhGAP guidelines for mental health sensitisation of community leaders were implemented in 1-day training events across 25 urban and rural health facilities (n = 1004 community leaders) in Uganda. A multiple choice mental health questionnaire was used to assess the community leaders’ mental health knowledge before and after completing the training. Training was evaluated across multiple sites and qualitative feedback comments were used to identify key themes on the impact of the training. The sensitisation training was found to be affordable, accessible and effective, and could be replicated in other LMICs and settings with local adaptations.

This study investigates the applicability of generative artificial intelligence (AI) in early-stage architectural design by evaluating the daylight performance of AI-generated sustainable housing plans across five distinct climate zones. A three-phase methodology was implemented: (1) Plan generation using text-to-image diffusion models (ChatGPT, Copilot, and LookX); (2) digital reconstruction in AutoCAD; and (3) daylight simulation via Velux Daylight Visualizer. Climate-adaptive prompts were formulated to guide the AI tools in producing context-specific floor plans with passive strategies. Out of 31 initial plans, eight valid outputs (five from ChatGPT and three from Copilot) were reconstructed in AutoCAD and simulated. Quantitative simulations were conducted on equinox and solstice dates, and average illuminance values were analyzed for key interior spaces (living room, kitchen, and bedroom). ChatGPT-generated plans demonstrated higher spatial clarity and more balanced daylight performance, whereas Copilot outputs varied significantly, and LookX was excluded due to insufficient architectural legibility. Results revealed that none of the models consistently integrated solar orientation or seasonal lighting considerations, indicating a gap between generative representation and environmental logic. The research contributes a replicable workflow that bridges generative AI and performance-based evaluation, offering critical insight into the current limitations and future potential of AI-assisted architectural design. The findings underscore the need for next-generation AI systems capable of semantic, spatial, and climatic reasoning to support environmentally responsive design practices.

Following the French example, the Meloni government has introduced the phrase ‘sovranità alimentare’ (sovereignty in food) into the title of the ministry of agriculture, and makes clear that it is engaging in a very determined effort to defend and promote the cultural heritage of Italian cuisine on all fronts, at home and abroad. But the origins of this impulse go back to the 1980s and the arrival of the McDonald’s hamburger chain, which gave birth to the Slow Food movement, now a global phenomenon. All this conceals several paradoxes: Italian cuisine has always been open to hybridised versions invented elsewhere (especially in America); production in key sectors, including wine, depends on large numbers of immigrant workers at a time when the government is trying to discourage immigration; and the ‘sovereignty in food’ concept unwittingly unites the government and some of its most radical opponents. But the very basis of this concept is challenged by the hyper-protectionist trade policy of the Trump administration.

This article examines the Philippines’ engagement with international law and institutions under Duterte’s populist presidency. While populism is often associated with hostility toward multilateralism, this case study reveals a more nuanced dynamic. The article argues that state engagement under populist administrations is more complex than assumed, and populist rhetoric does not uniformly dictate international behaviour. Using a novel conceptual framework and empirical data, it analyzes the Philippines’ multilateral interactions in human rights, trade, and health. Duterte’s government displayed ritualistic engagement with the United Nations Human Rights Council (UNHRC), constructive engagement with the World Trade Organization (WTO) and World Health Organization (WHO), and destructive disengagement from the International Criminal Court (ICC). Key themes include the divergence between rhetoric and action, instrumental use of institutions for domestic priorities, and the critical filtering role of domestic institutions. These findings offer broader insights as to how populist states balance domestic imperatives with international commitments, offering broader insights into the interplay between populism, foreign policy, and multilateralism.

Collaboration is both a process and an outcome. Collaboration is based on the idea that interactions between participants with a common goal, working together as partnerships and sharing resources, can solve complex or “wicked” problems that are not possible to solve in isolation. Collaboration may be simple, occurring between individuals, or more complex interorganizational arrangements across sectors, with the life cycle and size of the collaboration determined by the issue at hand. HTA collaborations may involve a wide range of stakeholders, including HTA agencies at the national, regional, or global level, academia, government (including regulatory authorities), industry, clinicians, providers, and patient organizations. Regardless of the number or type of participants, collaborations need a shared understanding of the common goal, an agreement on aims, and a commitment to shared solutions.

Industry and agency members of the Health Technology Assessment International (HTAi) Asia Policy Forum (APF) met in Seoul, South Korea, in November 2024 for open discussions on how to facilitate and improve the collaborative process between all stakeholders in the health system, including government, HTA agencies, industry, academia, clinicians, as well as patients. Over the three days, these discussions identified some of the risks and obstacles to collaboration in the region, how to develop and use collaboration better, as well as articulating the value and benefits of collaboration both in the region and globally.

This study provides a comprehensive analysis of the impact of helideck surface conditions on the safe operation of helicopter landing and take-off platforms on offshore drilling vessels. Over time, the deterioration of helideck surface coatings necessitates periodic friction coefficient testing every two years in compliance with international standards. Surface coatings that fail to meet the required thresholds are replaced, and the performance of the renewed surface is reassessed using the Helideck Micro GripTester (HMGT), in accordance with U.K. Safety Regulation Group CAP 437 (2023) standards for offshore helicopter landing areas. The findings indicate that the renewed helideck surface coatings lead to a significant increase in the coefficient of friction, thereby enhancing the stability of helicopters upon landing and while on deck. Independent sample t-test and correlation analyses confirmed statistically significant differences between the old and new surface conditions, demonstrating the positive impact of surface improvements on coefficient of friction and, therefore, operational safety. Furthermore, machine learning techniques were employed to model and analyse the non-linear relationships between surface conditions and flow number. The model results demonstrate that variations in helideck surface coatings directly influence helicopter performance and operational safety. These findings underscore the critical importance of regular resurfacing and friction testing in ensuring the safety and reliability of offshore helicopter operations.

Dyckia ibiramensis is an endemic species from southern Brazil, known for its ability to adapt to extreme environmental variations. This study reports the development of species-specific microsatellite markers and the assembly and annotation of the plastid genome of D. ibiramensis, aiming to generate new genomic resources useful for studies on the conservation and evolution of this endangered species. A total of 33,112 microsatellite loci were identified, of which 10 were selected and validated for genotyping 30 individuals from natural populations, showing high genetic variability. These 10 microsatellite markers were very informative for the evaluation of genetic variability. These microsatellite markers evidenced moderate to high genetic diversity at the individual level, low population differentiation, and the capacity of the species to recover population size from ancient genetic bottlenecks. The assembled plastid genome revealed conserved structures and the occurrence of features at the gene level, likely related to stress response to environmental conditions. This study expands the understanding of the genetics of D. ibiramensis, highlighting the importance of genomic strategies for the conservation of endangered species. In situ and ex situ conservation strategies should be used to avoid the extinction of this endemic southern Brazilian genetic resource in nature.

This article reflects on my experience recovering and organising two judicial archives in Uganda — the High Court and Mengo Court archives — which had long been neglected due to bureaucratic disinterest and legal hierarchies inherited from the colonial period. Together, these archives contained over 150,000 uncatalogued case files documenting how ordinary Ugandans used both native and colonial British courts to contest injustice and claim rights. Their exclusion from formal archives was not incidental; it reflected a legal order that considered British colonial statutes, appellate decisions and English-language records as authoritative expression of law, while dismissing native court records, as informal without precedential value and unworthy of preservation. Drawing on critical scholarship and my own archival experience, I argue that the archive of law should not be seen merely as a collection of sources or as sites of power and knowledge, but as a space where legal authority and legitimacy are produced, preserved, and erased. The recovery of these records not only unearthed a neglected legal history shaped by ordinary people but also challenges dominant narratives of legal authority in Africa. This work also exposed the omissions in the archival collections, the influence of colonial memory, and the importance of vernacular legal records for both historical research and contemporary legal practice.

This study examines how university curriculum reforms that increase course selection flexibility influence entrepreneurial outcomes. Departing from traditional emphasis on educational attainment, we explore how institutional changes in education shape entrepreneurial tendencies among alumni. Leveraging a reform that removed constraints on course selection at a major university, we find that increased educational choice significantly fosters entrepreneurship. Our analysis reveals partial support for the moderating effects of individual, family, and spatial factors: the positive impact of these reforms is contingent on the type of electives and courses taken, with stronger effects observed among alumni with entrepreneurial parents, those born in urban areas, and those from higher socioeconomic backgrounds. These groups leverage specific course patterns to align their educational choices with entrepreneurial aspirations, enhancing their likelihood of pursuing entrepreneurial ventures. However, these findings also underscore the potential for educational reforms to exacerbate inequalities, disproportionately benefiting those with preexisting advantages. By integrating insights on institutional changes, course-taking patterns, and individual moderators, this study advances understanding of the interplay between education and entrepreneurship, offering implications for designing more equitable educational policies.