Global food security worsened during the COVID-19 pandemic. In Nigeria, food security indicators increased in the first months of the pandemic and then decreased slightly but never returned to their pre-pandemic levels. We assess if savings groups provided household coping mechanisms during COVID-19 in Nigeria by combining the in-person LSMS-ISA/GHS-2018/19 with four rounds of the Nigerian Longitudinal Phone Survey collected during the first year of the pandemic. A quasi-difference-in-differences analysis setup leveraging the panel nature of the data indicates that savings group membership reduces the likelihood of skipping a meal but finds no statistically significant effect on the likelihood of running out of food or eating fewer kinds of food. Given theoretical priors and other literature positing a relationship, we also implement an OLS regression analysis controlling for baseline values finding that having at least one female household member in a savings group is associated with a 5–15% reduction in the likelihood of reporting skipping meals, running out of food, and eating fewer kinds of food. This analysis is not able to establish causality, however, and may in fact overestimate the effects. Together, the results indicate that savings group membership is positively associated with food security after COVID-19, but the causal effect is statistically significant for only one of the three food security indicators. To conclude, considering the interest in savings groups and expectations of continued food security shocks, the importance of collecting better gender-disaggregated longitudinal household data combined with experimental designs and institutional data on savings groups.

Heating, Ventilation, and Air Conditioning (HVAC) systems are major energy consumers in buildings, challenging the balance between efficiency and occupant comfort. While prior research explored generative AI for HVAC control in simulations, real-world validation remained scarce. This study addresses this gap by designing, deploying, and evaluating “Office-in-the-Loop,” a novel cyber-physical system leveraging generative AI within an operational office setting. Capitalizing on multimodal foundation models and Agentic AI, our system integrates real-time environmental sensor data (temperature, occupancy, etc.), occupants’ subjective thermal comfort feedback, and historical context as input prompts for the generative AI to dynamically predict optimal HVAC temperature setpoints. Extensive real-world experiments demonstrate significant energy savings (up to 47.92%) while simultaneously improving comfort (up to 26.36%) compared to baseline operation. Regression analysis confirmed the robustness of our approach against confounding variables like outdoor conditions and occupancy levels. Furthermore, we introduce Data-Driven Reasoning using Agentic AI, finding that prompting the AI for data-grounded rationales significantly enhances prediction stability and enables the inference of system dynamics and cost functions, bypassing the need for traditional reinforcement learning paradigms. This work bridges simulation and reality, showcasing generative AI’s potential for efficient, comfortable building environments and indicating future scalability to large systems like data centers.

In this editorial, we draw insights from a special collection of peer-reviewed papers investigating how new data sources and technology can enhance peace. The collection examines local and global practices that strive towards positive peace through the responsible use of frontier technologies. In particular, the articles of the collection illustrate how advanced techniques—including machine learning, network analysis, specialised text classifiers, and large-scale predictive analytics—can deepen our understanding of conflict dynamics by revealing subtle interdependencies and patterns. Others assess innovative approaches reinterpreting peace as a relational phenomenon. Collectively, they assess ethical, technical, and governance challenges while advocating balanced frameworks that ensure accountability alongside innovation. The collection offers a practical roadmap for integrating technical tools into peacebuilding to foster resilient societies and non-violent conflict transformations.

The aim of this study was to describe how the detection of protozoan and helminth parasites has been affected by the introduction of polymerase chain reaction (PCR) and changes in test algorithms. We extracted data about faecal samples tested for parasites (n = 114839) at five Norwegian clinical microbiology laboratories. Samples were classified into prePCR or postPCR depending on whether they were submitted before or after the introduction of PCR, and into diagnostic episodes (n = 99320). The number of diagnostic episodes increased 3.7-fold from prePCR to postPCR. Giardia positive episodes doubled, the positivity rate decreased from 2.0% to 1.3%. Cryptosporidium was hardly detected prePCR and increased to a positivity rate of 1.2%. Entamoeba histolytica was rarely found. Episodes examined for helminths decreased 51%, the number of positive episodes decreased 34%. Samples from immigrants were more likely to be positive for Giardia, E. histolytica, or helminths and less likely to be Cryptosporidium positive. During the COVID-19 pandemic, the number of Giardia and helminth-positive episodes decreased. Cryptosporidium-positive episodes remained unchanged. The implementation of multiplex PCR for protozoa led to a doubling of Giardia cases and a better test for Cryptosporidium. Fewer microscopy examinations raise concerns that helminth infections may be overlooked.

A random temporal graph is an Erdős-Rényi random graph $G(n,p)$, together with a random ordering of its edges. A path in the graph is called increasing if the edges on the path appear in increasing order. A set $S$ of vertices forms a temporal clique if for all $u,v \in S$, there is an increasing path from $u$ to $v$. Becker, Casteigts, Crescenzi, Kodric, Renken, Raskin and Zamaraev [(2023) Giant components in random temporal graphs. arXiv,2205.14888] proved that if $p=c\log n/n$ for $c\gt 1$, then, with high probability, there is a temporal clique of size $n-o(n)$. On the other hand, for $c\lt 1$, with high probability, the largest temporal clique is of size $o(n)$. In this note, we improve the latter bound by showing that, for $c\lt 1$, the largest temporal clique is of constant size with high probability.

The digital twin approach has gained recognition as a promising solution to the challenges faced by the Architecture, Engineering, Construction, Operations, and Management (AECOM) industries. However, its broader application across some AECOM sectors remains limited. A significant obstacle is that traditional DTs rely on deterministic models, which require deterministic input parameters. This limits their accuracy, as they do not account for the substantial uncertainties that are inherent in AECOM projects. These uncertainties are particularly pronounced in geotechnical design and construction. To address this challenge, we propose a probabilistic digital twin (PDT) framework that extends traditional DT methodologies by incorporating uncertainties and is tailored to the requirements of geotechnical design and construction. The PDT framework provides a structured approach to integrating all sources of uncertainty, including aleatoric, data, model, and prediction uncertainties, and propagates them throughout the entire modeling process. To ensure that site-specific conditions are accurately reflected as additional information is obtained, the PDT leverages Bayesian methods for model updating. The effectiveness of the PDT framework is showcased through an application to a highway foundation construction project, demonstrating its potential to integrate existing probabilistic methods to improve decision-making and project outcomes in the face of significant uncertainties. By embedding these methods within the PDT framework, we lower the barriers to practical implementation, making probabilistic approaches more accessible and applicable in real-world engineering workflows.

This paper documents the details of the design, verification, and certification of a novel technology: a remote monitoring system (digital twin) for a voyage data recorder, referred to as the HermAce Gateway. The electronic components, data transfer, and storage principle explain how the HermAce Gateway communicates and records safety-critical messages. Various prospective benefits to the industry are provided, primarily regarding the opportunities for remote support and testing that the digital twin facilitates. The HermAce Gateway was independently verified through a combination of semi-automated software in the loop and selected complimentary hardware in the loop tests. Different types of communication were simulated in multiple ways, including approximating real-world scenarios. Alarms contained in correctly formed messages were found to be detected and recorded by the HermAce Gateway, and a discussion of how this evidence can be quantified in the context of reducing uncertainty in the reliability of a digital twin. Certification of a digital system is a new concept in the maritime industry. The identification of functional requirements, which informed the verification testing, and the development of an AI register for what is expected to be an increasing number of such systems are also documented.

This article examines the impact of generative artificial intelligence (GAI) on higher education, emphasizing its effects in the broader educational contexts. As AI continues to reshape the landscape of teaching and learning, it is imperative for higher education institutions to adapt rapidly to equip graduates for the challenges of a progressively automated global workforce. However, a critical question emerges: will GAI lead to a more inclusive future of learning, or will it deepen existing divides and create a future where educational access and success are increasingly unequal? This study employs both theoretical and empirical approaches to explore the transformative potential of GAI. Drawing upon the literature on AI and education, we establish a framework that categorizes the essential knowledge and skills needed by graduates in the GAI era. This framework includes four key capability sets: AI ethics, AI literacy (focusing on human-replacement technologies), human–AI collaboration (emphasizing human augmentation), and human-distinctive capacities (highlighting unique human intelligence). Our empirical analysis involves scrutinizing GAI policy documents and the core curricula mandated for all graduates across leading Asian universities. Contrary to expectations of a uniform AI-driven educational transformation, our findings expose significant disparities in AI readiness and implementation among these institutions. These disparities, shaped by national and institutional specifics, are likely to exacerbate existing inequalities in educational outcomes, leading to divergent futures for individuals and universities alike in the age of GAI. Thus, this article not only maps the current landscape but also forecasts the widening educational gaps that GAI might engender.

During pregnancy, colonization by genital mycoplasmas may be associated with adverse outcomes. This study was conducted to investigate the prevalence of four species of Mollicutes (Mycoplasma hominis, Mycoplasma genitalium, Ureaplasma parvum, and Ureaplasma urealyticum) in pregnant women receiving high-risk prenatal care and to evaluate possible associated factors. Data collection included the application of a questionnaire and the collection of cervical swabs from pregnant women. Species identification was performed by real-time PCR. The overall prevalence of Mollicutes was 60.97%. 55.9% of pregnant women were colonized by Ureaplasma spp., and 19.51% by Mycoplasma spp. The prevalence rates by species were 48.78% for U. parvum, 11.59% for U. urealyticum, 18.9% for M. hominis, and 1.22% for M. genitalium. Age, 12 years of schooling or more, age at first sexual intercourse up to 14 years, third trimester of pregnancy, having undergone infertility treatment, presence of STI, and groin lymph nodes were associated with a higher prevalence of microorganisms. The results presented are of utmost importance for understanding the prevalence of these microorganisms, the characteristics of colonized pregnant women, and planning screening strategies and interventions that minimize the negative impacts of these infections.

In 1987, the United Nations Brundtland Commission defined sustainability as “meeting the needs of the present without compromising the ability of future generations to meet their own needs.” In recent years, the sustainability agenda has grown in importance, with many countries, regulators, industries shifting to implement sustainable practices. For retirement funds this means providing a lasting income in retirement for members, whilst ensuring a positive contribution to society and the environment. Retirement funds, with long-term liabilities, are therefore well placed and can play a significant role in contributing to the overall objective. This paper explores how retirement funds in various countries are progressing this agenda. We then introduce a sustainability reporting index, which measures the breadth and quality of how retirement funds can report on pricing in social and environmental externalities in the provision of a pension promise. The sustainability reporting index includes the financial inclusion aspects of retirement funds as well as how social and environmental externalities can be factored into the running of a fund and how its assets are invested. It explores the key areas that need to be monitored, the types of data required and the types of analytics that can be used by various stakeholders. The sustainability reporting index is intended to provide a benchmark against which various stakeholders can measure the effectiveness of their approach in pricing in these externalities. Actuaries of retirement funds can use the framework to go beyond focussing purely on the financial aspects of a fund, incorporating material non-financial aspects to ensure the provision of a sustainable pension income.

Invasive Group B Streptococcus (GBS) infection caused by the highly virulent Sequence Type 283 (ST283) strain has been linked to consumption of raw freshwater fish. In late summer 2024, enhanced surveillance in Hong Kong detected a surge of invasive ST283 cases.

A retrospective case–control study was conducted involving all invasive GBS patients reported during August to September 2024. Data were collected through standardised interviewer-administered questionnaires. Cases were defined as patients infected with the ST283 strain, while controls had non-ST283 strains. A multivariate logistic regression analysis was conducted to determine the risk factors.

Among 170 invasive GBS patients, 131 (77%) were identified as cases and 39 (23%) as controls. Physical handling of raw freshwater fish was found to be the strongest risk factor for ST283 infection (adjusted odds ratio: 8.4, 95% confidence interval: 1.4–50.1).

This study represents the first epidemiological evidence specifically linking physical contact with raw freshwater fish to an increased risk of invasive GBS ST283 infection. Effective interdepartmental coordination, intensive public health education, active surveillance, and prompt environmental interventions effectively mitigated this large outbreak. The findings underscore the need for sustainable preventive strategies targeting high-risk fish handling practices, particularly during warm periods favouring environmental proliferation of ST283.

The present paper develops a unified approach when dealing with short- or long-range dependent processes with finite or infinite variance. We are concerned with the convergence rate in the strong law of large numbers (SLLN). Our main result is a Marcinkiewicz–Zygmund law of large numbers for $S_{n}(f)= \sum_{i=1}^{n}f(X_{i})$, where $\{X_i\}_{i\geq 1}$ is a real stationary Gaussian sequence and $f(\!\cdot\!)$ is a measurable function. Key technical tools in the proofs are new maximal inequalities for partial sums, which may be useful in other problems. Our results are obtained by employing truncation alongside new maximal inequalities. The result can help to differentiate the effects of long memory and heavy tails on the convergence rate for limit theorems.

We consider a stochastic model, called the replicator coalescent, describing a system of blocks of k different types that undergo pairwise mergers at rates depending on the block types: with rate $C_{ij}\geq 0$ blocks of type i and j merge, resulting in a single block of type i. The replicator coalescent can be seen as a generalisation of Kingman’s coalescent death chain in a multi-type setting, although without an underpinning exchangeable partition structure. The name is derived from a remarkable connection between the instantaneous dynamics of this multi-type coalescent when issued from an arbitrarily large number of blocks, and the so-called replicator equations from evolutionary game theory. By dilating time arbitrarily close to zero, we see that initially, on coming down from infinity, the replicator coalescent behaves like the solution to a certain replicator equation. Thereafter, stochastic effects are felt and the process evolves more in the spirit of a multi-type death chain.

The primary focus of this article is to capture heterogeneous treatment effects measured by the conditional average treatment effect. A model averaging estimation scheme is proposed with multiple candidate linear regression models under heteroskedastic errors, and the properties of this scheme are explored analytically. First, it is shown that our proposal is asymptotically optimal in the sense of achieving the lowest possible squared error. Second, the convergence of the weights determined by our proposal is provided when at least one of the candidate models is correctly specified. Simulation results in comparison with several related existing methods favor our proposed method. The method is applied to a dataset from a labor skills training program.

In March 2024, the East Midlands Health Protection Team was notified of a case of invasive Group A Streptococcus (iGAS) infection in an elderly care home resident. Twenty-two days later, another case in a resident from the same floor of the care home was notified. In accordance with national guidelines, an outbreak was declared, and a multidisciplinary outbreak control team (OCT) was urgently convened. Screening for GAS throat carriage was undertaken for staff and residents, excluding those receiving end-of-life care. All isolates were strain typed and characterised. Infection prevention and control (IPC) visits were undertaken to provide ongoing support. Screening identified five residents and five staff members positive for GAS. Antibiotic prophylaxis was provided to all staff throughout the setting (n = 74) and all residents on the affected floor (n = 35). Three individuals were positive on repeat screening. All staff and residents screened negative after 4 months and the two clinical cases recovered. Eleven of the 12 GAS isolates were identified as emm 3.93. This outbreak highlighted the importance of rapid screening, possible only through the deployment of a dedicated team, and rescreening post-decolonising treatment, as a means to contain such outbreaks.

Lower COVID-19 vaccination coverage was observed among some populations with a migration background in the Netherlands. This study examined determinants of being unvaccinated against COVID-19 in the primary vaccination round in adults and in the 2022 autumn booster round in persons aged ≥60 years, among four populations of non-Dutch origin with below average vaccination coverage: Moroccan, Turkish, Surinamese and Dutch-Caribbean, and persons of Dutch origin. We performed a population-wide register-based study, examining associations between potential determinants and being unvaccinated using multivariable logistic regression and computing population attributable fractions. Being a migrant with two foreign-born parents, younger age, living in highly/extremely urban areas and having a lower income, lower education level and low medical risk for severe COVID-19 were risk factors for being unvaccinated in all populations. Substantial differences in the (strength of) determinants and population attributable fractions between populations were also observed. Socioeconomic status only partially mediated the association with being a migrant with two foreign-born parents. These findings illustrate that interventions targeting specific ethnic minority and migrant populations need further study with the aim to optimize the impact of vaccination programmes and improve health equity. To understand reasons behind non-uptake and design (community-based) interventions, qualitative and survey-based research is needed.

This paper has been prepared by the IFoA’s Collective Defined Contribution (CDC) working party. The purpose is to raise awareness within the actuarial community and pensions industry on the wide range of design options and considerations for CDC solutions, together with a set of principles for the design work, which we believe should apply in most cases. This should also aid understanding of why different designs are better in different circumstances, and why some designs might have certain features that others would avoid.

Determining the factors that impact the risk for infection with SARS-CoV-2 is a priority as the virus continues to infect people worldwide. The objective was to determine the effectiveness of vaccines and other factors associated with infection among Canadian healthcare workers (HCWs) followed from 15 June 2020 to 1 December 2023. We also investigate the association between antibodies to SARS-CoV-2 and subsequent infections with SARS-CoV-2. Of the 2474 eligible participants, 2133 (86%) were female, 33% were nurses, the median age was 41 years, and 99.3% had received at least two doses of COVID-19 vaccine by 31 December 2021. The incidence of SARS-CoV-2 was 0.91 per 1000 person-days. Prior to the circulation of the Omicron variants, vaccine effectiveness (VE) was estimated at 85% (95% CI 1, 98) for participants who received the primary series of vaccine. During the Omicron period, relative adjusted VE was 43% (95% CI 29, 54), 56% (95% CI 42, 67), and 46% (95% CI 24, 62) for 3, 4, and ≥ 5 doses compared with those who received primary series after adjusting for previous infection and other covariates. Exposure to infected household members, coworkers, or friends in the previous 14 days were risk factor for infection, while contact with an infected patient was not statistically significant. Participants with higher levels of immunoglobulin G (IgG) anti-receptor binding domain (RBD) antibodies had lower rates of infection than those with the lowest levels. COVID-19 vaccines remained effective throughout the follow-up of this cohort of highly vaccinated HCWs. IgG anti-RBD antibody levels may be useful as correlates of protection for issues such as vaccine development and testing. There remains a need to increase the awareness among HCWs about the risk of contracting SARS-CoV-2 from contacts at a variety of venues.