136 research outputs found
Исследование мирового опыта в нормировани инсоляции зданий и возможность интеграции Молдовы в эти процессы
Get PDFWhen designing buildings in dense urban areas, the problem of lack of sunlight in living rooms arises, which leads to increased energy costs for artificial lighting. The purpose of this study is to analyze the world experience and innovative solutions in the practice of regulating sunlight. This article study the technical regulations that regulate the insolation time in residential and public buildings in some countries of the world, the European Union, the CIS countries in comparison with the Republic of Moldova. It is proposed to improve the normative framework in this field for the Republic of Moldova
Large Minibatch Training on Supercomputers with Improved Accuracy and Reduced Time to Train
No full textLarge Cone Beam CT SCan Image Quality Improvement Using a Deep Learning U-Net Model
No full textCone beam CT scanners use much less radiation than to normal CT scans. However, compared to normal CT scans the images are noisy, showing several artifacts. The UNet Convolutional Neural Network may provide a way to reconstruct the a CT image from cone beam scans
Уровень фенотипической устойчивости и тип генных мутаций у M.tuberculosis
Get PDFThe adjustment of the minimum inhibitory concentrations for anti-tuberculosis drugs used in classical cultivation methods and the obtaining of more accurate data on the level of drug resistance, makes it possible to evaluate cases with different results to phenotypic and genotypic methods of anti-tuberculosis resistance testing. The evaluation of different types of genotypic resistance, with the description of mutations that confer low or high resistance, is consistent with the level of phenotypic resistance, and also makes it possible to adjust treatment regimens, which ultimately will positively influence the duration and results of treatment. Whenever testing of M. tuberculosis resistance by molecular methods allows, the results should be reported with the specific mutations detected and the description of the clinical implications of this mutationAjustarea concentraţiilor minime inhibitorii pentru preparatele antituberculoase utilizate în metodele de cultură clasice şi obţinerea unor date mai exacte despre nivelul de rezistenţă la medicamente, face posibilă evaluarea cazurilor cu rezultate diferite la metodele fenotipice şi genice de testare a rezistenţei antituberculoase. Evaluarea diferitor tipuri de rezistenţă genică, cu descrierea mutaţiilor care conferă rezistenţă joasă sau înaltă, este în coerenţă nivelul de rezistenţă fenotipică, şi de asemenea face posibilă ajustarea schemelor de tratament, care la rândul lor influenţează pozitiv la durata şi rezultatele tratamentului. Ori de câte ori testarea rezistenţei M. tuberculosis prin metode moleculare permite, rezultatele trebuie să fie raportate cu mutaţiile specifice detectate şi cu descrierea implicaţiilor clinice ale prezentei mutaţieiРегулировка минимальных ингибирующих концентраций противотуберкулезных препаратов, используемых в классических методах культивирования, и получение более точных данных об уровне лекарственной устойчивости, позволяет оценивать случаи с разными результатами фенотипических и генотипических методов противотуберкулезной устойчивости. Оценка различных типов генотипической устойчивости с описанием мутаций, придающих низкую или высокую устойчивость, согласуется с уровнем фенотипической устойчивости, а также позволяет корректировать схемы лечения, что в конечном итоге положительно влияет на продолжительность и результаты лечение. Если позволяет тестирование устойчивости M. tuberculosis молекулярными методами, результаты должны быть сообщены с указанием конкретных обнаруженных мутаций и описанием клинических последствий этой мутаци
Learning to Evaluate Chess Positions with Deep Neural Networks and Limited Lookahead
Get PDFIn this paper we propose a novel supervised learning approach for training Artificial Neural Networks (ANNs) to evaluate chess positions. The method that we present aims to train different ANN architectures to understand chess positions similarly to how highly rated human players do. We investigate the capabilities that ANNs have when it comes to pattern recognition, an ability that distinguishes chess grandmasters from more amateur players. We collect around 3,000,000 different chess positions played by highly skilled chess players and label them with the evaluation function of Stockfish, one of the strongest existing chess engines. We create 4 different datasets from scratch that are used for different classification and regression experiments. The results show how relatively simple Multilayer Perceptrons (MLPs) outperform Convolutional Neural Networks (CNNs) in all the experiments that we have performed. We also investigate two different board representations, the first one representing if a piece is present on the board or not, and the second one in which we assign a numerical value to the piece according to its strength. Our results show how the latter input representation influences the performances of the ANNs negatively in almost all experiments
Compact amorphous-silicon visible-light monitor integrated in silicon nitride waveguides
No full textThis work reports on the realization of an amorphous silicon visible-light detector integrated in Si3N4 waveguides. The device is very compact (< 40 μm), has a responsivity of about 10 mA/W and a sensitivity of -40 dBm.Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.QID/Hanson LabQN/Quantum Nanoscienc
- …
