Recent advances in AutoML have led to automated tools that can compete with machine learning experts on supervised learning tasks. In this work, we present two versions of Auto-Net, which provide automatically-tuned deep neural networks without any human intervention. The first version, Auto-Net 1.0, builds upon ideas from the competition-winning system Auto-sklearn by using the Bayesian Optimization method SMAC and uses Lasagne as the underlying deep learning (DL) library. The more recent Auto-Net 2.0 builds upon a recent combination of Bayesian Optimization and HyperBand, called BOHB, and uses PyTorch as DL library. To the best of our knowledge, Auto-Net 1.0 was the first automatically-tuned neural network to win competition datasets against human experts (as part of the first AutoML challenge). Further empirical results show that ensembling Auto-Net 1.0 with Auto-sklearn can perform better than either approach alone, and that Auto-Net 2.0 can perform better yet.

There is growing interest in blood eosinophil counts in the management of chronic respiratory conditions such as asthma and chronic obstructive pulmonary disease (COPD). Despite this, typical blood eosinophil levels in the general population, and the impact of potential confounders on these levels have not been clearly defined.We measured blood eosinophil counts in a random sample of 11 042 subjects recruited from the general population in Austria. We then: 1) identified factors associated with high blood eosinophil counts (>75th percentile); and 2) excluded subjects with these factors to estimate median blood eosinophil counts in a “healthy” sub-population (n=3641).We found that: 1) in the entire cohort, age ≤18 years (OR 2.41), asthma (OR 2.05), current smoking (OR 1.72), positive skin prick test (OR 1.64), COPD (OR 1.56), metabolic syndrome (OR 1.41), male sex (OR 1.36) and obesity (OR 1.16) were significantly (p<0.05) associated with high blood eosinophil counts (binary multivariable logistic regression analysis), and had an additive effect; and 2) after excluding these factors, in those older than 18 years, blood eosinophil counts were higher in males than in females (median 120 (5%–95% CI: 30–330) versus 100 (30–310) cells·µL−1, respectively) and did not change with age.Median blood eosinophil counts in adults are considerably lower than those currently regarded as normal, do not change with age beyond puberty, but are significantly influenced by a variety of factors which have an additive effect. These observations will contribute to the interpretation of blood eosinophil levels in clinical practice.

This paper describes a computerized alternative to glottochronology for estimating elapsed time since parent languages diverged into daughter languages. The method, developed by the Automated Similarity Judgment Program (ASJP) consortium, is different from glottochronology in four major respects: (1) it is automated and thus is more objective, (2) it applies a uniform analytical approach to a single database of worldwide languages, (3) it is based on lexical similarity as determined from Levenshtein (edit) distances rather than on cognate percentages, and (4) it provides a formula for date calculation that mathematically recognizes the lexical heterogeneity of individual languages, including parent languages just before their breakup into daughter languages. Automated judgments of lexical similarity for groups of related languages are calibrated with historical, epigraphic, and archaeological divergence dates for 52 language groups. The discrepancies between estimated and calibration dates are found to be on average 29% as large as the estimated dates themselves, a figure that does not differ significantly among language families. As a resource for further research that may require dates of known level of accuracy, we offer a list of ASJP time depths for nearly all the world's recognized language families and for many subfamilies. The greater the degree of linguistic differentiation within a stock, the greater is the period of time that must be assumed for the development of such differentiations.

Background The Lung, hEart, sociAl, boDy (LEAD) Study (ClinicalTrials.gov; NCT01727518; http://clinicaltrials.gov ) is a longitudinal, observational, population-based Austrian cohort that aims to investigate the relationship between genetic, environmental, social, developmental and ageing factors influencing respiratory health and comorbidities through life. The general working hypothesis of LEAD is the interaction of these genetic, environmental and socioeconomic factors influences lung development and ageing, the risk of occurrence of several non-communicable diseases (respiratory, cardiovascular, metabolic and neurologic), as well as their phenotypic (ie, clinical) presentation. Methods LEAD invited from 2011–2016 a random sample (stratified by age, gender, residential area) of Vienna inhabitants (urban cohort) and all the inhabitants of six villages from Lower Austria (rural cohort). Participants will be followed-up every four years. A number of investigations and measurements were obtained in each of the four domains of the study (Lung, hEart, sociAl, boDy) including data to screen for lung, cardiovascular and metabolic diseases, osteoporosis, and cognitive function. Blood and urine samples are stored in a biobank for future investigations. Results A total of 11.423 males (47.6%) and females (52.4%), aged 6–80 years have been included in the cohort. Compared to governmental statistics, the external validity of LEAD with respect to age, gender, citizenship, and smoking status was high. Conclusions In conclusion, the LEAD cohort has been established following high quality standards; it is representative of the Austrian population and offers a platform to understand lung development and ageing as a key mechanism of human health both in early and late adulthood.