Backyard chickens are increasingly popular, and their husbandry varies widely. How backyard chickens are housed may influence the accessibility of chicken feed and water to wild birds, and thus, the contact rates between both groups. Increased contacts have implications for pathogen transmission; for instance, Newcastle disease virus or avian influenza virus may be transmitted to and from backyard chickens from contaminated water or feed. Given this potentially increased pathogen risk to wild birds and backyard chickens, we examined which wild bird species are likely to encounter backyard chickens and their resources. We performed a supplemental feeding experiment followed by observations at three sites associated with backyard chickens in North Georgia, USA. At each site, we identified the species of wild birds that: (a) shared habitat with the chickens, (b) had a higher frequency of detection relative to other species and (c) encountered the coops. We identified 14 wild bird species that entered the coops to consume supplemental feed and were considered high-risk for pathogen transmission. Our results provide evidence that contact between wild birds and backyard chickens is frequent and more common than previously believed, which has crucial epidemiological implications for wildlife managers and backyard chicken owners.

Fraud analytics refers to the use of advanced analytics (data mining, big data analysis, or artificial intelligence) to detect fraud. While fraud analytics offers the promise of more efficiency in fighting fraud, it also raises legal challenges related to data protection and administrative law. These legal requirements are well documented but the concrete way in which public administrations have integrated them remains unexplored. Due to the complexity of the techniques applied, it is crucial to understand the current state of practice and the accompanying challenges to develop appropriate governance mechanisms. The use of advanced analytics in organizations without appropriate organizational change can lead to ethical challenges and privacy issues. The goal of this article is to examine how these legal requirements are addressed in public administrations and to identify the challenges that emerge in doing so. For this, we examined two case studies related to fraud analytics from the Belgian Federal administration: the detection of tax frauds and social security infringements. This article details 15 governance practices that have been used in administrations. Furthermore, it highlights the complexity of integrating legal requirements with advanced analytics by identifying six key trade-offs between fraud analytics opportunities and legal requirements.

Contact with livestock and consumption of unpasteurised dairy products are associated with an increased risk of zoonotic and foodborne infection, particularly among populations with close animal contact, including pastoralists and semi-pastoralists. However, there are limited data on disease risk factors among pastoralists and other populations where livestock herding, particularly of dromedary camels, is common. This cross-sectional study used a previously validated survey instrument to identify risk factors for self-reported symptoms. Adults (n = 304) were randomly selected from households (n = 171) in the Somali Region of Ethiopia, a region characterised by chronic food insecurity, population displacement, recurrent droughts and large semi-pastoralist and pastoralist populations. Multivariable logistic regression assessed associations between self-reported symptoms and type of milk consumed, controlling for demographics and human-animal interaction. Consumption of days-old unrefrigerated raw camel milk was significantly associated with symptoms in the 30 days prior to the survey (AOR = 5.07; 95% CI 2.41–10.66), after controlling for age, refugee status, sanitation, camel ownership and source of drinking water and accounting for clustering. Consumption of days-old unrefrigerated raw ruminant milk was significantly associated with symptoms (AOR = 4.00, 95% CI 1.27–12.58). Source of drinking water and camel ownership, a proxy for camel contact, were significantly associated with the outcome in each model. There were no significant associations between self-reported symptoms and fresh or soured animal milk consumption. Research is needed to identify pathogens and major routes of transmission. Tailored communication campaigns to encourage safe food preparation should also be considered.

For a uniform random labelled tree, we find the limiting distribution of tree parameters which are stable (in some sense) with respect to local perturbations of the tree structure. The proof is based on the martingale central limit theorem and the Aldous–Broder algorithm. In particular, our general result implies the asymptotic normality of the number of occurrences of any given small pattern and the asymptotic log-normality of the number of automorphisms.

Reproducibility of a deep-learning fully convolutional neural network is evaluated by training several times the same network on identical conditions (database, hyperparameters, and hardware) with nondeterministic graphics processing unit operations. The network is trained to model three typical time–space-evolving physical systems in two dimensions: heat, Burgers’, and wave equations. The behavior of the networks is evaluated on both recursive and nonrecursive tasks. Significant changes in models’ properties (weights and feature fields) are observed. When tested on various benchmarks, these models systematically return estimations with a high level of deviation, especially for the recurrent analysis which strongly amplifies variability due to the nondeterminism. Trainings performed with double floating-point precision provide slightly better estimations and a significant reduction of the variability of both the network parameters and its testing error range.