Performance of gender detection tools: a comparative study of name-to-gender inference services

Objective: To evaluate the performance of gender detection tools that allow the uploading of files (e.g., Excel or CSV files) containing first names, are usable by researchers without advanced computer skills, and are at least partially free of charge.Methods: The study was conducted using four physician datasets (total number of physicians: 6,131; 50.3% female) from Switzerland, a multilingual country. Four gender detection tools met the inclusion criteria: three partially free (Gender API, NamSor, and genderize.io) and one completely free (Wiki-Gendersort). For each tool, we recorded the number of correct classifications (i.e., correct gender assigned to a name), misclassifications (i.e., wrong gender assigned to a name), and nonclassifications (i.e., no gender assigned). We computed three metrics: the proportion of misclassifications excluding nonclassifications (errorCodedWithoutNA), the proportion of nonclassifications (naCoded), and the proportion of misclassifications and nonclassifications (errorCoded).Results: The proportion of misclassifications was low for all four gender detection tools (errorCodedWithoutNA between 1.5 and 2.2%). By contrast, the proportion of unrecognized names (naCoded) varied: 0% for NamSor, 0.3% for Gender API, 4.5% for Wiki-Gendersort, and 16.4% for genderize.io. Using errorCoded, which penalizes both types of error equally, we obtained the following results: Gender API 1.8%, NamSor 2.0%, Wiki-Gendersort 6.6%, and genderize.io 17.7%.Conclusions: Gender API and NamSor were the most accurate tools. Genderize.io led to a high number of nonclassifications. Wiki-Gendersort may be a good compromise for researchers wishing to use a completely free tool. Other studies would be useful to evaluate the performance of these tools in other populations (e.g., Asian).

Enhancing the Interpretability of Deep Models in Healthcare Through Attention: Application to Glucose Forecasting for Diabetic People

The adoption of deep learning in healthcare is hindered by their “black box” nature. In this paper, we explore the RETAIN architecture for the task of glucose forecasting for diabetic people. By using a two-level attention mechanism, the recurrent-neural-network-based RETAIN model is interpretable. We evaluate the RETAIN model on the type-2 IDIAB and the type-1 OhioT1DM datasets by comparing its statistical and clinical performances against two deep models and three models based on decision trees. We show that the RETAIN model offers a very good compromise between accuracy and interpretability, being almost as accurate as the LSTM and FCN models while remaining interpretable. We show the usefulness of its interpretable nature by analyzing the contribution of each variable to the final prediction. It revealed that signal values older than 1[Formula: see text]h are not used by the RETAIN model for 30[Formula: see text]min ahead of time prediction of glucose. Also, we show how the RETAIN model changes its behavior upon the arrival of an event such as carbohydrate intakes or insulin infusions. In particular, it showed that the patient’s state before the event is particularly important for the prediction. Overall the RETAIN model, thanks to its interpretability, seems to be a very promising model for regression or classification tasks in healthcare.

Left Hemiclamshell Approach for Posterior Aortic Arch Aneurysm

The anatomical situation of posterior aortic arch aneurysms is a surgical challenge. The surgical approach should not only guarantee a safe dissection of the supra-aortic trunks and the descending aorta but also allow the cannulation for extracorporeal circulation. Indeed, protecting the cerebral flow is essential. Another challenge is to preserve both the phrenic and recurrent nerves while sparing chest wall muscles. A hemiclamshell approach for posterior aortic arch aneurysm seems to be a good compromise regarding these requirements.

Policies and practice in the disclosure of medical error: Insights from leading countries to address the issue in Italy

The relationship between physician and patient has undergone profound changes in recent years. Patients increasingly insist on being thoroughly informed with detailed information about treatments and procedures suggested for their best care. This is also due to the growing suspicion towards doctors and the health-care system in general. Therefore, it is no longer possible to hide a medical error. To satisfy the request for honesty and safety of patients and society, it is necessary to enhance the skills and tools that physicians can use when disclosing and explaining an error to the patient. All modern codes of medical conduct acknowledge the importance of strengthening communication between physician and patient, which is the only way to save a relationship under constant threat of rupture and to improve the quality and safety of the treatment. The disclosure and explanation of the error has become not only an ethical duty but also a prudent way of avoiding negligence lawsuits. In this context, in 2013, Germany approved a law known as Patientenrechtegesetz, which we consider a good compromise between patient expectations and the need for doctors to work without the constant fear of being sued for malpractice. This work seeks to provide an overview of the most important issues pertaining to disclosure of medical error and of practice in other countries, with the aim of offering a contribution to the debate on this subject in Italy.

Optimal void finders in weak lensing maps

ABSTRACT Cosmic voids are a key component of the large-scale structure that contain a plethora of cosmological information. Typically, voids are identified from the underlying galaxy distribution, which is a biased tracer of the total matter field. Previous works have shown that 2D voids identified in weak lensing (WL) maps – WL voids – correspond better to true underdense regions along the line of sight. In this work, we study how the properties of WL voids depend on the choice of void finder, by adapting several popular void finders. We present and discuss the differences between identifying voids directly in the convergence maps, and in the distribution of WL peaks. Particular effort has been made to test how these results are affected by galaxy shape noise (GSN), which is a dominant source of noise in WL observations. By studying the signal-to-noise ratios (S/N) for the tangential shear profile of each void finder, we find that voids identified directly in the convergence maps have the highest S/N but are also the ones most affected by GSN. Troughs are least affected by noise, but also have the lowest S/N. The tunnel algorithm, which identifies voids in the distribution of WL peaks, represents a good compromise between finding a large tangential shear S/N and mitigating the effect of GSN.

What Is the Minimum Volume of Sample to Find Small Microplastics: Laboratory Experiments and Sampling of Aveiro Lagoon and Vouga River, Portugal

Small microplastics (<1 mm) comprise a great fraction of microplastics (<5 mm) found in the environment and are often overlooked due to the constrains of transporting and filtering large volumes of water in grab samplings. The objective of this work was to determine the minimum volume for reliable quantification of small microplastics in the environment. Different volumes (0.1, 0.25, 0.5, 1, 2.5 L) of laboratory spikes (fresh and saltwater) and environmental samples were filtered. Sampling volumes of 0.5 L or 1 L are a good compromise between drawbacks, such as effort, time, organic and mineral matter, potential contamination, and reliability of results, evaluated by interquartile range, accuracy, coefficient of variation, and recovery rates. Moreover, the observation of Nile Red-stained environmental samples under 470 nm produced six-times higher concentrations than samples under 254 nm, namely, 18 microplastics L−1 and 3 microplastics L−1 for the Aveiro Lagoon and 1 microplastics L−1 and 0 microplastics L−1 for the Vouga River, Portugal. This work also raises concerns about the underreporting of environmental concentrations of microplastics.

Using Evolutionary Algorithms to Achieve Sustainable Solutions through Household Appliances

In order to reach a sustainable planet, there is a permanent need by the consumer (decision- agent), to achieve sustainable solutions, with its decisions. Given the importance of the buildings, as a sector to achieve such solutions, as well as the diversity of household appliances existent on the market, with all its different issues, there are several tradeoffs to consider (e.g. energy and water consumption vs initial investment), which difficult the consumer’s choices from the market. The problem increases, since nowadays, the consumer tries to get a solution from the market, with a good compromise between the Economic, Social and Environmental dimensions, and according to its specific needs, which can be different from other consumers. By considering a multicriteria approach, combined with an optimization technique, based on Evolutionary Algorithms (EA), it’s provided a set of sustainable solutions from the market to the consumer that respects the compromise referred before. In this work, it is presented an approach to support a decision-agent (DA) (consumer), by performing a set of sustainable choices based on household appliances from the market and suitable to its needs. Based on the obtained solutions, several savings are achieved (electrical and water consumption, CO2 emissions), by considering the consumer’s relative importance, regarding its Economics, Environmental and Social concerns.