75 research outputs found
Electrically tunable non-radiative lifetime in WS2/WSe2 heterostructures
No full textVan der Waals heterostructures based on transition metal dichalcogenides (TMDs) have emerged as excellent candidates for next-generation optoelectronics and valleytronics, due to their fascinating physical properties. The understanding and active control of the relaxation dynamics of heterostructures play a crucial role in device design and optimization. Here, we investigate the back-gate modulation of exciton dynamics in a WS2/WSe2 heterostructure by combining time-resolved photoluminescence (TRPL) and transient absorption spectroscopy (TAS) at cryogenic temperatures. We find that the non-radiative relaxation lifetimes of photocarriers in heterostructures can be electrically controlled for samples with different twist-angles, whereas such lifetime tuning is not present in standalone monolayers. We attribute such an observation to doping-controlled competition between interlayer and intralayer recombination pathways in high-quality WS2/WSe2 samples. The simultaneous measurement of TRPL and TAS lifetimes within the same sample provides additional insight into the influence of coexisting excitons and background carriers on the photo-response, and points to the potential of tailoring light-matter interactions in TMD heterostructures
Teaching chinese as a second foreign language: communication and culture
No full textИ Аньжань, аспирант/ ассистент, Институт иностранного языка, Российский университет
дружбы народов, Москва, Россия.
Yi Anran, graduate student/ assistant, Institute of Foreign Languages, Peoples' Friendship University
of Russia, Moscow, Russia.В настоящее время китаеведение стремительно развивается во всем мире. Большое
количество иностранцев изучают китайский язык. Ученые отмечают неразрывную связь между обществом, культурой, историей, географией и языком. Поэтому коммуникация является сложным
процессом, требующим тщательного и сознательного погружения в контекст языка. Целью статьи
является исследование особенностей обучения китайскому языку как второму иностранному. Задачи: выявить специфику развития коммуникативной компетентности учащихся, охарактеризовать
сущность и структуру коммуникативной компетентности, включающей в себя социальные и культурные знания, проанализировать внутриязыковые и внеязыковые элементы обучения китайскому
языку как второму иностранному в контексте межкультурной коммуникации. В статье использовались методы: сравнение, ретроспективный, системно-логический анализ. Автором дана характеристика коммуникативных стратегий при изучении китайского языка, а также рассмотрены вопросы,
касающиеся их выбора во взаимосвязи с различными факторами. Обращено внимание на весомую
роль коммуникативных задач, которые являются ключевым элементом обучения китайскому языку,
а также на экспериментальную проверку эффективности модели профессионально-ориентированного
обучения китайскому языку студентов гуманитарных специальностей в российских вузах. Результаты исследования показывают, что внеучебная деятельность повышает мотивацию студентов к изучению китайского языка, уровень самоорганизации и ответственности. Грамотно выстроенная модель профессионально-ориентированного обучения, включающая большой объём внеучебной деятельности, является достаточно эффективной в плане формирования межкультурной компетенции,
необходимой для дальнейшей работы выпускников российских вузов. At present, Sinology is rapidly developing all over the world. A large number of foreigners
are learning Chinese. Scientists note the inextricable link between society, culture, history, geography and
language. In this regard, communication is a complex process that requires careful and conscious immersion
in the language context. The purpose of the article is to study the features of teaching Chinese as a second
foreign language. The objectives of the study are to identify the specifics of the development of students'
communicative competence, to characterize the essence and structure of communicative competence, which
includes social and cultural knowledge, to analyze intralinguistic and extralinguistic elements of teaching
Chinese as a second foreign language in the context of intercultural communication. The following methods
were used in the article: comparison, retrospective analysis and system-logical analysis. The author gives a description of communicative strategies in the study of the Chinese language, and also considers issues related
to their choice in relation to various factors. The focus of the research are communicative tasks which
are a key element of teaching the Chinese language and the experimental verification of the effectiveness of
the model of professionally oriented teaching of the Chinese language to students of the humanities in Russian
universities. The results of the study show that extracurricular activities increase students' motivation to
learn Chinese, increase the level of self-organization and responsibility. A well-built model of professionally
oriented education, which includes a large amount of extracurricular activities, is effective in terms of the formation
of intercultural competence necessary for the further work of graduates of Russian universities
Retrieval and Personalised Segmentation of Amateur Free-Hand Sketches
Get PDFSketch-based 3D model retrieval focus on retrieving relevant 3D models using sketch(es) as input. It is attractive to use sketch in category-level 3D retrieval because sketches by hand provide an easy way to input, yet they are detailed enough to specify shapes. Compared to category-level sketch-based 3D retrieval, instance-level retrieval gives user more pragmatic retrieval results. The users draw the sketch, and the system will yield the exact 3D model that is desired. However, as the number of 3D models is limited in the dataset, there might not exist the exact 3D model that the user wants. But, we could get a similar alternative by assembling suitable 3D parts in the current dataset. This underlines the need for part-level sketch-based 3D retrieval. In this task, one pre- requirement is that sketches need to be segmented by parts. This emphasises the need for practical semantic sketch segmentation.In this thesis, we explore the problem of sketch-based 3D retrieval in a coarse-to-fine manner, i.e., from category-level to instance-level to part-level retrieval. And show that sketch is especially powerful when communicating with the high-dimensional 3D world, where currently humans still have difficulty in interpreting its visuals and understand- ing how interaction occurs. Besides, to facilitate the part-level sketch-based 3D shape retrieval, we address the problem of personalised sketch segmentation for the first time.For all the aforementioned tasks, we perform rigorous experiments on various publicly available datasets. The quantitative and qualitative results are included to demonstrate the effectiveness of our proposed solutions
Современные методики обучения китайскому языку студентов гуманитарных специальностей в российских вузах
No full textCurrently, there is a huge demand for Chinese and specialists who speak the language at the highest level, which leads to the creation of new teaching methods, new developments and a more detailed study of existing ones. This article presents the results of research on the methodology of teaching Chinese to students of humanities of Russian universities. The article aims to orient the degree of elaboration of the issues of teaching the Chinese language. The purpose of this work is to use modern teaching methods that can increase motivation to study and keep students interested in Chinese languages. The main results of the study are the systematization of the main mechanisms, principles of using modern methods in Chinese language classes. The material of the article can be useful for teachers and methodologists of the Chinese language. This publication examines the approaches of various researchers to the development of teaching methods to.В настоящее время существует огромный спрос на китайский язык и специалистов, владеющих языком на самом высоком уровне, что приводит к созданию новых методик преподавания, новых разработок и более детальному изучению существующих. В данной статье представлены результаты исследования по методике преподавания китайского языка студентам гуманитарных специальностей российских университетов. Статья призвана сориентировать степень разработанности вопросов преподавания китайского языка. Целью данной работы является использование современных методов обучения, которые могут повысить мотивацию к учебе и поддержать интерес студентов к китайским языкам. Основными результатами исследования являются систематизация основных механизмов, принципов использования современных методов на занятиях по китайскому языку. Материал статьи может быть полезен преподавателям и методистам китайского языка. В данной публикации рассматриваются подходы различных исследователей к разработке методов преподавания китайского языка
TEACHING CHINESE AS A FOREIGN LANGUAGE AT UNIVERSITIES IN RUSSIA AND CHINA
No full textThe processes of globalization and integration taking place in the world, the growing interaction between countries with different cultures, increasingly require the training of personnel with the skills and knowledge necessary to solve the problems of globalization and integration in foreign languages. To a greater extent, in the modern world, there is no need for interpreters to act as intermediaries between specialists and professionals from various fields with the necessary language skills. In recent decades, in Russia, due to the active development of contacts, interest in studying Oriental languages, especially Chinese, has been steadily growing. The Institute of Foreign Languages (IFL) of the Peoples' Friendship University of Russia (RUDN University) offers the opportunity to study Chinese as a foreign language. The authors of the article give a comparative analysis of the teaching of foreign language methods in Russian and Chinese universities to create a communicative-active approach to teaching Chinese as a foreign language. They pointed out the following teaching principles: gradual and consistent mastering of material, repetition, testing theory by practice, the principle of organizing interactive classes and self-study. Methodical recommendations are offered for educators who teach Chinese as a foreign language in the form of complex exercises and language games. The article aims to choose the best educational trajectory for students from Intralinguodidactics. Based on personal experience in IFL RUDN University and a comparative analysis of the methods of teaching Chinese in Russia and China, the authors share their methodical tips in FLT. Special attention is paid to lexical and hieroglyphic games. The game manages to achieve a shift in an angle: a language is no longer an object of study, but a specific tool, a tool that it should be in the natural environment. Students understand that language can be used. Features of teaching Chinese as a second foreign language determine the originality of the techniques and methods used by the teacher. A relatively small number of class hours, the lack of multidimensionality, and limited opportunities for language practice require careful and methodically correct organization of classroom lessons, specific instructions for independent work of students and strict control over it
A CASE STUDY: DEVELOPING COMPETENCES OF A FOREIGN LANGUAGE WITH TUIS (TTIS, TELECOMMUNICATION TRAINING AND INFORMATION SYSTEM) IN THE RUDN UNIVERSITY
No full textIntroduction to Python programming in finance
No full textThe main purpose of this project is to develop an efficacious teaching programme for finance students to master Python programming and enhance their learning experiences.
The author designed two complete projects usings contents from CZ1103 Introduction to Computational Thinking and Programming as these projects can be used to complement the module curriculum. These two projects, split into 6 sections each, aim to strengthen students’ knowledge as well as educate them on how to break down problems before analyzing and solving them. Through this, students can gain hands-on practice on programming as well as gain greater confidence and interest.Bachelor of Science in Data Science and Artificial Intelligenc
Direct evidence of microbiological water quality changes on bacterial quantity and community caused by plumbing system
No full textDrinking water quality deteriorates from treatment plant to customer taps, especially in the plumbing system. There is no direct evidence about what the differences are contributed by plumbing system. This study compared the water quality in the water main and at customer tap by preparing a sampling tap on the water main. The biomass was quantified by adenosine triphosphate (ATP) and the microbial community was profiled by 454 pyrosequencing. The results showed that in distribution pipes, biofilm contributed >94% of the total biomass, while loose deposits showed little contribution (< 2%) because of the low amount of loose deposits. The distribution of biological stable water had minor effects on the microbiocidal water quality regarding both quantity (ATP 1 ng/L vs. 1.7 ng/L) and community of the bacteria. Whereas the plumbing system has significant contribution to the increase of active biomass (1.7 ng/L vs. 2.9 ng/L) and the changes of bacterial community. The relative abundance of Sphingomonas spp. at tap (22%) was higher than that at water main (2%), while the relative abundance of Pseudomonas spp. in tap water (15%) was lower than that in the water from street water main (29%). Though only one location was prepared and studied, the present study showed that the protocol of making sampling tap on water main offered directly evidences about the impacts of plumbing system on tap water quality, which makes it possible to distinguish and study the processes in distribution system and plumbing system separately.Sanitary Engineerin
Artificial intelligence clinical support system trained with spectral domain optical coherence tomography volumetric data for glaucoma assessment
No full textPh.D.Glaucoma is a major cause of irreversible visual morbidity worldwide, and it is projected to affect 111.8 million people by 2040. Early detection of glaucoma is essential for prompt treatment, vision preservation and blindness prevention. However, the early disease stages are usually asymptomatic, and thus glaucoma patients are often at the late stage when they present to ophthalmologists. At present, opportunistic case screening in eye clinics is the primary approach for glaucoma detection, but this fails to reach the vast majority of the population at large and leads to nearly half of glaucoma patients remaining undiagnosed. Population-based screening for glaucoma is currently not cost-effective nor practical, due to a shortage of highly trained workforce and lack of efficient testing modality. Community-based glaucoma screening among high-risk individuals has been suggested to be a more cost-effective and efficient approach. However, a large number of highly trained personnel and expertise are still required to be on-site for interpreting the results or making any referral decisions. Glaucoma screening has, therefore, remained costly, inefficient, time-consuming, and labour-intensive, and hard to fulfil in clinics without glaucoma experts.Spectral-domain optical coherence tomography (SD-OCT) is a non-contact and non-invasive imaging technology offering three-dimensional (3D), objective, and quantitative assessment of the retinal layers, including the retinal nerve fibre layer (RNFL). Comparison of measured RNFL thickness (RNFLT) with a built-in normative database in SD-OCT has become a widely used assessment for glaucoma detection and monitoring in the last decade. However, there is no specific threshold for diagnosing or excluding glaucoma purely from RNFLT in the conventional SD-OCT. Moreover, RNFLT measurement is affected by various factors, such as poor image quality or myopia. Hence, experienced glaucoma specialists or highly trained human graders are needed to assess image quality and the metrics for glaucoma identification.In addition to assessing structural changes in RNFL and optic nerve head (ONH), functional assessment based on the visual field (VF) testing is perceived as essential. However, effective and reliable VF assessment could be difficult, and it would be impractical to conduct VF testing individually for glaucoma screening among the population.Artificial intelligence (AI) has shown great potential in disease screening, classification, and management. Deep learning (DL), under the broad umbrella of AI, has proven to be a promising clinical support tool. Several studies have applied DL for referable glaucomatous optic neuropathy (GON) detection from two-dimensional (2D) fundus photographs or quantitative metrics generated by SD-OCT, as well as prediction of VF metrics. These attempts demonstrated the efficiency and accuracy of DL algorithms in glaucoma detection. However, there are several remaining gaps in such application. First, there was no image quality control incorporated with glaucoma-detection DL algorithms and the algorithms only trained on gradable images. Indeed, initial gradability assessment is essential in any clinical workflows as this would involve image re-acquisition, referral pathway or subsequent glaucoma assessment. Second, DL algorithms trained with 2D fundus photographs only focused on glaucoma structural changes from the top view of retina and ONH, and they did not consider the layered retina or the entire ONH structure. Third, fundus morphological changes, including peripapillary atrophy (PPA), optic disc tilting and fundus tessellation, could influence the classification of GON. For example, PPA and fundus tessellation result in general RNFL thinning, and the degree of vertical optic disc tilting is associated with a more temporally positioned RNFL thickness peak. Previous studies seldom considered these morphological changes for DL model development and therefore resulted in misclassified output. Fourth, glaucoma-related structural changes are correlated with VF loss. Recent studies showed that global VF indices could be predicted from RNFL thickness maps or en face SD-OCT images. However, the RNFL thickness maps highly relied on accurate segmentation, and the en face images only illustrated the top view of areas around the ONH. Moreover, the macular area, a region corresponding to the visual sensitivity distribution, was not included in some of the reported studies for the prediction. Finally, there was a lack of validation with external datasets in most of these studies. External validation or testing is indeed crucial to verify the generalizability of a DL model and to show its potential application in real-world settings.This thesis aimed to 1) introduce the background, objectives and hypothesis; 2) describe overall materials and methods; 3) develop and validate a 3D DL model based on SD-OCT volumetric scans for automated image quality control; 4) develop and validate a 3D DL model to detect GON using SD-OCT volumetric scans; 5) develop and validate a multi-task 3D DL model based on SD-OCT volumetric scans to discriminate GON and fundus morphological changes simultaneously; 6) develop and validate a 3D DL model to predict global VF indices from SD-OCT volumetric scans; 7) and finally discuss the strengths and limitations of all studies, as well as future directions. Ultimately, this thesis aimed to deliver an integrated AI clinical support system using state-of-the-art DL techniques for glaucoma assessment with SD-OCT.In the first study, we used squeeze-and-excitation residual network next dimension (SE-ResNeXt) architectures to build a 3D DL model for automated image quality control. SD-OCT volumetric optic disc scans obtained from a commercial device (Carl Zeiss Meditec, Dublin, CA, USA) were labelled as ungradable when signal strength (SS) was 25% of the peripheral area. Artifacts included: 1) blurriness, 2) off-centration, 3) poor axial alignment, 4) signal loss, 5) motion artifacts, and 6) mirror artefact. SD-OCT volumetric scans were labelled as gradable when SS was ≥5, and absence of artifacts in the RNFL measurement circle, and/or presence of artifacts in only 25％外圍區域時，從商用設備（Carl Zeiss Meditec, Dublin, CA, USA）獲得的SD-OCT立體視盤掃描被標記為不可分級。偽影包括（1）偏離中心，（2）未對準，（3）信號丟失，（4）運動偽像，（5）鏡像偽像和（6）模糊。當信號強度大於等於5且沒有任何偽影或在僅<25％的外圍區域中存在偽影而視網膜神經纖維層測量圈中沒有時，SD-OCT立體掃描被標記為可讀。我們的發現表明，在內部和外部驗證中，該深度學習模型可以自動過濾不可分讀立體掃描，並具有良好表現。在第二項研究中，我們開發了基於殘差網絡（ResNet）結構的三維深度學習模型以實現自動青光眼檢測。根據可讀SD-OCT圖像上的視網膜神經纖維層變薄和具有相應位置視野缺損的標準，將SD-OCT立體視盤掃描標記為有或沒有青光眼。生成的熱點圖顯示，用於檢測青光眼的三維深度學習模型所學習的特徵與臨床醫生所觀察的特徵相似。在內部和外部驗證中，該三維深度學習模型的表現可與經驗豐富的青光眼專家相媲美。在第三項研究中，我們使用SD-OCT立體視盤掃描訓練了一個三維多任務ResNet深度學習模型，以同時檢測青光眼以及眼底或視神經頭周圍區域的形態變化。眼底形態變化定義為存在視乳頭周圍萎縮，視盤傾斜和眼底豹紋狀改變。據我們所知，該多任務深度學習是首次嘗試從SD-OCT立體視盤掃描中同時檢測青光眼和眼底形態變化。它具有更好的通用性，並且在內部和外部驗證中均達到了與使用RNFL厚度的傳統分類方法相當的表現。在最後一項研究中，我們基於已有的捲積神經網絡開發並驗證了預測性三維深度學習模型，以學習青光眼結構變化與相應功能變化之間的關係。青光眼患者的配對SD-OCT立體視盤掃描，黃斑掃描和視野報告用於訓練，測試和驗證該深度學習模型。預測視野和真實視野綜合指標之間的一致性很好，綜合指標包括視野指數，平均偏差和圖案標準偏差。尤其是在早期青光眼中，平均偏差的預測和真實值一致性很好。這將有可能減少視野測試這一繁瑣且相對主觀過程的頻率，並減輕診所的負擔。總而言之，上述三維深度學習模型的良好性能證明了人工智能有望通過實現自動圖像質量控制，結構變化檢測和功能變化預測，將SD-OCT設備作為用於青光眼評估的單一研究工具。未來必須進行進一步的前瞻性驗證和增量成本效益分析，以證明這些深度學習模型在臨床中的實用性。最終，無論是在初級，二級或者三級臨床水平，該整合的基於人工智能深度學習的臨床支持系統可以潛在地應用到青光眼人群篩查計劃以及臨床治療流程中，以提高青光眼檢測和監測的效率，準確性和成本效益。更重要的是，這一系統具有很大潛力，通過遠程醫療服務來輔助缺乏資源的地區進行青光眼治療。Ran, Anran.Thesis Ph.D. Chinese University of Hong Kong 2020.Includes bibliographical references (leaves 148-162).Abstracts also in Chinese.Title from PDF title page (viewed on March 22, 2022)
- …
