Possibility of increasing the switching resource of liquid-metal self-regulating fuses with fusion insert

THE PURPOSE. Assessment of the state of the art in the development of current limiters based on liquid metal self-healing fuses and an increase in the switching life of self-healing fuses with a composite fuse-link.METHODS. When solving the problem, the method of literature analysis was used, as well as the method for calculating dependencies for different materials of the fuse-link, the implemented method of scientific computer mathematics.RESULTS. Literary sources contain information about the designs of liquid-metal self-healing fuses and their research, allowing conclusions to be drawn about their viability and the possibility of obtaining a practical yield. A common disadvantage of the of liquid-metal self-healing fuses design is a limited switching resource due to an increase in the diameter of the fuse-link under the action of arc erosion. A significant effect of increasing the switching resource and the stability of the protective characteristic is possessed by the design of a of liquid-metal self-healing fuses with a composite fusible link. However, in such designs, the switching capacity is reduced. It is shown that the main criterion for the operability of the liquid-metal self-healing fuses with a composite fusible link is the value of the ratio between the value of the shunting resistance and the resistance of the electric ARC. The performance of the liquid-metal self-healing fuses with a composite fusible link is ensured at values of this value below critical.CONCLUSION. The proposed mathematical expression may turn out to be useful in the development of liquid-metal self-healing fuses with a composite fusible link design, for example, when choosing an electrode material, liquid metal, etc. This will significantly increase the switching resource of the liquid-metal self-healing fuses with a composite fusible link. The condition must be taken into account in the parametric synthesis of the current limiter based on the liquid-metal self-healing fuses with a composite fusible link.

Transfer, storage and processing of large volumes of scientific data

The textbook examines large scientific projects and the amount of data generated by them, and provides an overview of scientific computer networks that allow high-speed transmission of large amounts of data for these projects. The article considers the computing systems offered by the leading manufacturers of computer equipment for processing large amounts of data and providing both the possibilities of storing large amounts of data, including distributed data, and the means of analytics and parallel data processing in real time. Special attention is paid to the security of the transmitted scientific information. Meets the requirements of the federal state educational standards of higher education of the latest generation. For students of technical specialties of bachelor's degree, master's degree, specialty, studying in the areas of "Applied Mathematics and Computer Science", "Business Computer Science" and "Computer Science and Computer Engineering".

DeepEthogram: a machine learning pipeline for supervised behavior classification from raw pixels

Researchers commonly acquire videos of animal behavior and quantify the prevalence of behaviors of interest to study nervous system function, the effects of gene mutations, and the efficacy of pharmacological therapies. This analysis is typically performed manually and is therefore immensely time consuming, often limited to a small number of behaviors, and variable across researchers. Here, we created DeepEthogram: software that takes raw pixel values of videos as input and uses machine learning to output an ethogram, the set of user-defined behaviors of interest present in each frame of a video. We used convolutional neural network models that compute motion in a video, extract features from motion and single frames, and classify these features into behaviors. These models classified behaviors with greater than 90% accuracy on single frames in videos of flies and mice, matching expert-level human performance. The models accurately predicted even extremely rare behaviors, required little training data, and generalized to new videos and subjects. DeepEthogram runs rapidly on common scientific computer hardware and has a graphical user interface that does not require programming by the end-user. We anticipate DeepEthogram will enable the rapid, automated, and reproducible assignment of behavior labels to every frame of a video, thus accelerating all those studies that quantify behaviors of interest.Code is available at: https://github.com/jbohnslav/deepethogram

Code Sharing in Psychological Methods and Statistics: An Overview and Associations with Conventional and Alternative Research Metrics

Towards discovering more effective means to promote code sharing as an open science practice, this study explores the current state of code sharing within the field of psychological methods and statistics and examines the association between this practice and conventional and alternative research metrics. Towards this purpose, a total of 815 articles from three major journals within the field of psychological methods and statistics were manually screened and encoded based on code sharing practices and general article characteristics. In addition, data on conventional (citation counts) and alternative research metrics (Altmetric Attention Score) was retrieved through online scientific databases (Web of Science and Altmetric.com). This input was then analysed using descriptive statistics and regression models suitable for count data, and robustness of the findings were assessed using multiverse analysis. The findings of this study suggest that the sharing of scientific computer code is not (yet) extensively practiced within the field of psychological methods and statistics. In the majority of academic articles included in this study, scientific computer code was not shared (66 %). Moreover, if such code was shared, it was frequently found to be improperly annotated (70 %) and/or incomplete (52 %). Nevertheless, the findings of this study also suggest a hopeful prospect, as the sharing of scientific computer code has increased between 2010 and 2017. The study revealed a robust positive connection with alternative research metrics. This study did not find robust positive connection between code sharing and citation counts.

Virtual Realism: Really Realism or only Virtually so? A Comment on D. J. Chalmers’s Petrus Hispanus Lectures

What is the status of a cat in a virtual reality environment? Is it a real object? Or part of a fiction? Virtual realism, as defended by D. J. Chalmers, takes it to be a virtual object that really exists, that has properties and is involved in real events. His preferred specification of virtual realism identifies the cat with a digital object. The project of this paper is to use a comparison between virtual reality environments and scientific computer simulations to critically engage with Chalmers’s position. I first argue that, if it is sound, his virtual realism should also be applied to objects that figure in scientific computer simulations, e.g. to simulated galaxies. This leads to a slippery slope because it implies an unreasonable proliferation of digital objects. A philosophical analysis of scientific computer simulations suggests an alternative picture: The cat and the galaxies are parts of fictional models for which the computer provides model descriptions. This result motivates a deeper analysis of the way in which Chalmers builds up his realism. I argue that he buys realism too cheap. For instance, he does not really specify what virtual objects are supposed to be. As a result, rhetoric aside, his virtual realism isn’t far from a sort of fictionalism.

A research institution framework for publishing open code to enable reproducible science

Reproducible science is greatly aided by open publishing of scientific computer code. There are also many institutional benefits for encouraging the publication of scientific code, but there are also institutional considerations around intellectual property and risk. We discuss questions around scientific code publishing from the perspective of a research organisation asking: who will be involved, how should code be licensed, where should code be published, how to get credit, what standards, and what costs? In reviewing advice and evidence relevant to these questions we propose a research institution framework for publishing open scientific code to enable reproducible science.

A research institution framework for publishing open code to enable reproducible science

Reproducible science is greatly aided by open publishing of scientific computer code. There are also many institutional benefits for encouraging the publication of scientific code, but there are also institutional considerations around intellectual property and risk. We discuss questions around scientific code publishing from the perspective of a research organisation asking: who will be involved, how should code be licensed, where should code be published, how to get credit, what standards, and what costs? In reviewing advice and evidence relevant to these questions we propose a research institution framework for publishing open scientific code to enable reproducible science.

These Models Need Enterprise Data Management!

After 28 years of public service, the Information Technology Bureau Chief for the Southwest Florida Water Management District (SWFWMD), Dr. Steven Dicks, was still passionate about serving the public. By first focusing on the needs of people in the District, Dr. Dicks was able to objectively observe the many aspects of operations for which he was responsible. His observations led him to believe that the use of technology within his organization wasn’t as effective and efficient as he knew it could be. The SWFWMD managed the water resources of 16 Florida counties in an area that was close to 10,000 square miles, with 4.7 million inhabitants. To effectively manage water resources, the organization needed to understand current water needs, prepare for future needs, and protect and preserve water resources within its boundaries. A significant part of managing water resources involved the production and utilization of scientific computer models that help track, predict, and control a plethora of water related challenges. At the SWFWMD, individuals used desktop computers to run most of the scientific models. Given the limited computing capacity of the average desktop computer, running a model was beginning to require too much time; and simple interruptions, such as a system reboot, could jeopardize the ability to complete long model runs. Therefore, the resulting data used and produced was often very ineffectively managed. Dr. Dicks began to wonder about the current system of producing and managing scientific models, and how the systems might be improved. He evaluated options. He thought upgrading individual desktops could help, but only in the short term because as the data sets and demand for processing power continued to grow, the desktop might always be a step behind. Installing powerful servers to house the data sets and run the models would be a significant improvement, but the cost to acquire and maintain the new system might challenge the budget. A cloud-based solution utilizing an “infrastructure as a service” approach, was a third option, but current system infrastructure compatibility, security, and access needed to be carefully evaluated. Dr. Dicks believed that data management was just as important as the processing power of an upgrade, and that ultimately the technology that allowed for the most effective EDM system needed to be identified and implemented to best serve the needs of the District into the future.