Isabel Agadagba, Triphonia Kilasara, Takudzwa Tarutira
et al.
Digital payment systems have become a cornerstone of consumer finance in Africa. Prominent payment categories include money transfer applications, mobile money, cryptocurrencies, stablecoins, and central bank digital currencies (CBDCs). While there are studies exploring how and why people use individual digital payment systems (both in Africa and beyond), we lack a good understanding of why people choose between different categories of payment systems, and how they view the tradeoffs between different categories. We conducted qualitative interviews in three African countries -- Nigeria, Tanzania, and Zimbabwe -- to understand how and why people use various payment systems, and what influenced them to start using these systems. Our study highlights several notable findings regarding tradeoffs between perceived utility, privacy, and security. For example, many users trust government issuers to protect them from scams, but they do not trust those same institutions to build reliable systems and products or prioritize customer satisfaction. We also find that most users have accounts on multiple payment systems, and conduct a complex selection process using different platforms for different types of payments. This selection process is driven in part by financial considerations, but also by security, privacy, and trust preferences. Our findings suggest compelling directions for regulators and the research community to design systems that balance users' trust and utility needs.
Samantha Biegel, David Brüggemann, Francesco Grossi
et al.
Natural and anthropogenic disturbances are impacting the health of forests worldwide. Monitoring forest disturbances at scale is important to inform conservation efforts. Here, we present a scalable approach for country-wide mapping of forest greenness anomalies at the 10 m resolution of Sentinel-2. Using relevant ecological and topographical context and an established representation of the vegetation cycle, we learn a predictive quantile model of the normalised difference vegetation index (NDVI) derived from Sentinel-2 data. The resulting expected seasonal cycles are used to detect NDVI anomalies across Switzerland between April 2017 and August 2025. Goodness-of-fit evaluations show that the conditional model explains 65% of the observed variations in the median seasonal cycle. The model consistently benefits from the local context information, particularly during the green-up period. The approach produces coherent spatial anomaly patterns and enables country-wide quantification of forest browning. Case studies with independent reference data from known events illustrate that the model reliably detects different types of disturbances.
This article explores the overlooked involvement of women in supporting the University of Leuven during the sixteenth and seventeenth centuries. Despite gendered regulations against the presence of women in early modern universities, archival evidence reveals their significant roles in providing scholarships to students. Through detailed analysis of testamentary records, the study uncovers women’s deep engagement with the university’s activities. Spanning various social backgrounds, these women demonstrated a profound understanding of the university’s operations and influenced its development through financial support and by designing instructions for student conduct and academic focus. Despite often being sidelined in historical narratives, their contributions challenge traditional views of male-dominated academia and underscore the diverse roles women played in shaping early modern universities.
Negative campaigning is a central feature of political competition, yet empirical research has been limited by the high cost and limited scalability of existing classification methods. This study makes two key contributions. First, it introduces zero-shot Large Language Models (LLMs) as a novel approach for cross-lingual classification of negative campaigning. Using benchmark datasets in ten languages, we demonstrate that LLMs achieve performance on par with native-speaking human coders and outperform conventional supervised machine learning approaches. Second, we leverage this novel method to conduct the largest cross-national study of negative campaigning to date, analyzing 18 million tweets posted by parliamentarians in 19 European countries between 2017 and 2022. The results reveal consistent cross-national patterns: governing parties are less likely to use negative messaging, while ideologically extreme and populist parties -- particularly those on the radical right -- engage in significantly higher levels of negativity. These findings advance our understanding of how party-level characteristics shape strategic communication in multiparty systems. More broadly, the study demonstrates the potential of LLMs to enable scalable, transparent, and replicable research in political communication across linguistic and cultural contexts.
In this paper, we present TalkingMachines -- an efficient framework that transforms pretrained video generation models into real-time, audio-driven character animators. TalkingMachines enables natural conversational experiences by integrating an audio large language model (LLM) with our video generation foundation model. Our primary contributions include: (1) We adapt a pretrained SOTA image-to-video DiT into an audio-driven avatar generation model of 18 billion parameters; (2) We enable infinite video streaming without error accumulation through asymmetric knowledge distillation from a bidirectional teacher model into a sparse causal, autoregressive student model; (3) We design a high-throughput, low-latency inference pipeline incorporating several key engineering optimizations such as: (a) disaggregation of the DiT and VAE decoder across separate devices, (b) efficient overlap of inter-device communication and computation using CUDA streams, (c) elimination of redundant recomputations to maximize frame-generation throughput. Please see demo videos here - https://aaxwaz.github.io/TalkingMachines/
Starting from the assumption that printing privileges were an important way of enhancing a publication’s reputation, this article focuses on psalters, which played an important role in Protestant religious life in the Dutch Republic. Psalters were among the earliest books granted privileges by the States of Holland and the States-General, and were published with a privilege more often than other works. Furthermore, while privileges were generally applied for by publishers, in the case of psalters it was often the psalmists themselves who were the applicant. This article argues that this remarkable engagement of psalmists and printers of psalters in the system of printing privileges interacted with the pluralism of the seventeenth-century Dutch religious landscape, showing how the contexts of the privileged psalters diversified: whereas the first privileges were connected to the Dutch Revolt and the creation of a strong Reformed church, later on in the seventeenth century privileged psalters also became important within other churches. An analysis of the use of the privileges in the front matter of psalters suggests that the sense of political approval of the privilege interacted with the religious approval that psalmists sought. When aiming at an official position within the church, a psalmist was probably at a disadvantage if their work was lacking such a privilege.
N. Y. Al Saiqal, James. C. Ryabn, Osiris Jorge Parcero
The United Arab Emirates (UAE) is a young, oil-rich country, where national youth display a clear preference for public sector employment. Growing youth unemployment reinforces the importance of non-government employment, including entrepreneurship. This study investigates UAE national youth intentions towards entrepreneurship through the Theory of Planned Behavior and the Entrepreneurial Intention Questionnaire (EIQ). Analysis (N=544) identifies the direct influence of attitude and perceived behavioral control, and indirect influence of subjective norms on entrepreneurship intention. Results also examine several demographic variables and highlight the potential importance of family and social groups in promoting entrepreneurial intentions in this emerging country context.
In this paper, we demonstrated the benefit of using pre-trained model to extract acoustic embedding to jointly predict (multitask learning) three tasks: emotion, age, and native country. The pre-trained model was trained with wav2vec 2.0 large robust model on the speech emotion corpus. The emotion and age tasks were regression problems, while country prediction was a classification task. A single harmonic mean from three metrics was used to evaluate the performance of multitask learning. The classifier was a linear network with two independent layers and shared layers, including the output layers. This study explores multitask learning on different acoustic features (including the acoustic embedding extracted from a model trained on an affective speech dataset), seed numbers, batch sizes, and normalizations for predicting paralinguistic information from speech.
Phosphorus (P) is considered to be one of the key elements for life, making it an important element to look for in the abundance analysis of spectra of stellar systems. Yet, there exists only a handful of spectroscopic studies to estimate the P abundances and investigate its trend across a range of metallicities. We have observed full HK band spectra at a spectral resolving power of R=45,000 with IGRINS instrument. Abundances are determined using SME in combination with 1D MARCS stellar atmosphere models. The investigated sample of stars have reliable stellar parameters estimated using optical FIES spectra (GILD; Jönsson et al. in prep.). In order to determine the P abundances from the 16482.92 Angstrom P line, we take special care of the CO($ν=7-4$) blend. We determine the C, N, O abundances from atomic carbon and a range of non-blended molecular lines (CO, CN, OH) which are aplenty in the H band region of K giant stars, assuring an appropriate modelling of the blending CO($ν=7-4$) line. We present [P/Fe] vs [Fe/H] trend for 38 K giant stars in the metallicity range of -1.2 dex $<$ [Fe/H] $<$ 0.4 dex. We find that our trend matches well with the compiled literature sample of prominently dwarf stars and limited number of giant stars. Our trend is found to be higher by $\sim$ 0.05 - 0.1 dex compared to the theoretical chemical evolution trend in Cescutti et al. 2012 resulting from core collapse supernova (type II) of massive stars with the P yields from Kobayashi et al. (2006) arbitrarily increased by a factor of 2.75. Thus the enhancement factor might need to be $\sim$ 0.05 - 0.1 dex higher to match our trend. We also find an empirically determined primary behaviour for phosphorus. Furthermore, the phosphorus abundance is found to be elevated by $\sim$ 0.6 - 0.9 dex in two metal poor s-enriched stars compared to the theoretical chemical evolution trend.
During an ongoing epidemic, especially in the case of a new agent, data are partial and sparse, also affected by external factors as for example climatic effects or preparedness and response capability of healthcare structures. Despite that we showed how, under some universality assumptions, it is possible to extract strategic insights by modelling the pandemic trough a probabilistic Polya urn scheme. In the Polya framework, we provide both the distribution of infected cases and the asymptotic estimation of the incidence rate, showing that data are consistent with a general underlying process at different scales. Using European confirmed cases and diagnostic test data on COVID-19 , we provided an extensive comparison among European countries and between Europe and Italy at regional scale, for both the two big waves of infection. We globally estimated an incidence rate in accordance with previous studies. On the other hand, this quantity could play a crucial role as a proxy variable for an unbiased estimation of the real incidence rate, including symptomatic and asymptomatic cases.
Saurabh Bansal, Jeong Han Kim, Christopher Kolda
et al.
The mirror twin Higgs model (MTH) is a solution to the Higgs hierarchy problem that provides well-predicted cosmological signatures with only three extra parameters: the temperature of the twin sector, the abundance of twin baryons, and the vacuum expectation value (VEV) of twin electroweak symmetry breaking. These parameters specify the behavior of twin radiation and the acoustic oscillations of twin baryons, which lead to testable effects on the cosmic microwave background (CMB) and large-scale structure (LSS). While collider searches can only probe the twin VEV, through a fit to cosmological data we show that the existing CMB (Planck18 TTTEEE+lowE+lowT+lensing) and LSS (KV450) data already provide useful constraints on the remaining MTH parameters. Additionally, we show that the presence of twin radiation in this model can raise the Hubble constant $H_0$ while the scattering twin baryons can reduce the matter fluctuations $S_8$, which helps to relax the observed $H_0$ and $S_8$ tensions simultaneously. This scenario is different from the typical $Λ$CDM + $ΔN_{\rm eff}$ model, in which extra radiation helps with the Hubble tension but worsens the $S_8$ tension. For instance, when including the SH0ES and Planck SZ data in the fit, we find that a universe with $\gtrsim 20\%$ of the dark matter comprised of twin baryons is preferred over $Λ$CDM by $\sim4σ$. If the twin sector is indeed responsible for resolving the $H_0$ and $S_8$ tensions, future measurements from the Euclid satellite and CMB Stage 4 experiment will further measure the twin parameters to $O(1-10\%)$-level precision. Our study demonstrates how models with hidden naturalness can potentially be probed using precision cosmological data.
Abdellah El Mekki, Abdelkader El Mahdaouy, Kabil Essefar
et al.
Dialect and standard language identification are crucial tasks for many Arabic natural language processing applications. In this paper, we present our deep learning-based system, submitted to the second NADI shared task for country-level and province-level identification of Modern Standard Arabic (MSA) and Dialectal Arabic (DA). The system is based on an end-to-end deep Multi-Task Learning (MTL) model to tackle both country-level and province-level MSA/DA identification. The latter MTL model consists of a shared Bidirectional Encoder Representation Transformers (BERT) encoder, two task-specific attention layers, and two classifiers. Our key idea is to leverage both the task-discriminative and the inter-task shared features for country and province MSA/DA identification. The obtained results show that our MTL model outperforms single-task models on most subtasks.
Precipitated by rapid globalization, rising inequality, population growth, and longevity gains, social protection programs have been on the rise in low- and middle-income countries (LMICs) in the last three decades. However, the introduction of public benefits could displace informal mechanisms for risk-protection, which are especially prevalent in LMICs. If the displacement of private transfers is considerably large, the expansion of social protection programs could even lead to social welfare loss. In this paper, we critically survey the recent empirical literature on crowd-out effects in response to public policies, specifically in the context of LMICs. We review and synthesize patterns from the behavioral response to various types of social protection programs. Furthermore, we specifically examine for heterogeneous treatment effects by important socioeconomic characteristics. We conclude by drawing on lessons from our synthesis of studies. If poverty reduction objectives are considered, along with careful program targeting that accounts for potential crowd-out effects, there may well be a net social gain.
A previous study of symmetric collisions of massive nuclei has shown that current models of multi-nucleon transfer (MNT) reactions do not adequately describe the transfer product yields. To gain further insight into this problem, we have measured the yields of MNT products in the interaction of 977 (E/A = 4.79 MeV) and 1143 MeV (E/A = 5.60 MeV) $^{204}$Hg with $^{208}$Pb. We find that the yield of multi-nucleon transfer products are similar in these two reactions and are substantially lower than those observed in the reaction of 1257 MeV (E/A = 6.16 MeV) $^{204}$Hg + $^{198}$Pt. We compare our measurements with the predictions of the GRAZING-F, di-nuclear systems (DNS) and improved quantum molecular dynamics (ImQMD) models. For the observed isotopes of the elements Au, Hg, Tl, Pb and Bi, the measured values of the MNT cross sections are orders of magnitude larger than the predicted values. Furthermore, the various models predict the formation of nuclides near the N=126 shell, which are not observed.
Stefan Hochwarter, Do Duy Cuong, Nguyen Thi Kim Chuc
et al.
Previous studies have shown that health information technologies have a positive impact on health systems. Electronic health record (EHR) systems are one of the most promising applications, demonstrating a positive effect in high income countries. On the other hand, robust evidence for low and middle income countries is still spare. The aim of this study is to initiate a carefully planned nationwide EHR system in Vietnam by assessing the core readiness. The assessment structure is mainly based on previous research, which recommends a readiness assessment prior to to an EHR system implementation. To collect data, participant observation, document analysis and an in-depth interview were used. This study has revealed new insights into the current situation on EHR in Vietnam. The Ministry of Health is currently working on improving the conditions for future implementation of a Vietnamese EHR system. There are issues with the current way of handling health records. These issues are encouraging the Ministry of Health to work on identifying the next steps for an EHR system implementation. The integration of an EHR system with current systems seems to be challenging as most systems are commercial, closed source and very likely have no standardised interface. In conclusion, this study identifies points which need to be further investigated prior to an implementation. Generally, health care workers show good awareness of new technologies. As the Vietnam's health care system is centrally organised, there is the possibility for a nation-wide implementation. This could have a positive impact on the health care system, however, besides rigours planning also standards need to be followed and common interfaces implemented. Finally, this assessment has focused on only one level of readiness assessment. Further research is needed to complete the assessment.
Combinatorial/probabilistic models for cross-country dual-meets are proposed. The first model assumes that all runners are equally likely to finish in any possible order. The second model assumes that each team is selected from a large identically distributed population of potential runners and with each potential runner's ranking determined by the initial draw from the combined population.
We present a low-space overhead simulation algorithm based on the truncated Dyson series for time-dependent quantum dynamics. This algorithm is applied to simulating time-independent Hamiltonians by transitioning to the interaction picture, where some portions are made time-dependent. This can provide a favorable complexity trade-off as the algorithm scales exponentially better with derivatives of the time-dependent component than the original Hamiltonian. We show that this leads to an exponential improvement in gate complexity for simulating some classes of diagonally dominant Hamiltonian. Additionally we show that this can reduce the gate-complexity scaling for simulating $N$-site Hubbard models for time $t$ with arbitrary long-range interactions as well as reduce the cost of quantum chemistry simulations within a similar-sized plane-wave basis to $\widetilde{\mathcal{O}}(N^2t)$ from $\widetilde{\mathcal{O}}(N^{11/3}t)$. We also show a quadratic improvement in query complexity for simulating sparse time-dependent Hamiltonians, which may be of independent interest.