Data scarcity is a major challenge when training deep learning (DL) models. DL demands a large amount of data to achieve exceptional performance. Unfortunately, many applications have small or inadequate data to train DL frameworks. Usually, manual labeling is needed to provide labeled data, which typically involves human annotators with a vast background of knowledge. This annotation process is costly, time-consuming, and error-prone. Usually, every DL framework is fed by a significant amount of labeled data to automatically learn representations. Ultimately, a larger amount of data would generate a better DL model and its performance is also application dependent. This issue is the main barrier for many applications dismissing the use of DL. Having sufficient data is the first step toward any successful and trustworthy DL application. This paper presents a holistic survey on state-of-the-art techniques to deal with training DL models to overcome three challenges including small, imbalanced datasets, and lack of generalization. This survey starts by listing the learning techniques. Next, the types of DL architectures are introduced. After that, state-of-the-art solutions to address the issue of lack of training data are listed, such as Transfer Learning (TL), Self-Supervised Learning (SSL), Generative Adversarial Networks (GANs), Model Architecture (MA), Physics-Informed Neural Network (PINN), and Deep Synthetic Minority Oversampling Technique (DeepSMOTE). Then, these solutions were followed by some related tips about data acquisition needed prior to training purposes, as well as recommendations for ensuring the trustworthiness of the training dataset. The survey ends with a list of applications that suffer from data scarcity, several alternatives are proposed in order to generate more data in each application including Electromagnetic Imaging (EMI), Civil Structural Health Monitoring, Medical imaging, Meteorology, Wireless Communications, Fluid Mechanics, Microelectromechanical system, and Cybersecurity. To the best of the authors’ knowledge, this is the first review that offers a comprehensive overview on strategies to tackle data scarcity in DL.
ABSTRACT
The researcher seeks to shed light on the relationship analysis and the impact between organizational values in all its dimensions (Administration Management, Mission, relationship management, environmental management) and strategic performance (financial perspective, customer perspective, the perspective of internal processes, learning and development) in the presidency of Two Universities of Baghdad & Al-Nahrain, it has been formulating three hypotheses for this purpose.
The main research problem has been the following question: Is there a relationship and the impact of bet
... Show MoreThe article aims to consider the concept of language metaphor in Russian and Arabic languages and the problem of metaphor functioning in language, since it is one of the most important figurative components of the structural organization of the text and an important means of reflecting the national culture of each people. and often in revealing the image of a metaphor one can feel the full flexibility of the language and its beauty.
The process of soil classification in Iraq for industrial purposes is important topics that need to be extensive and specialized studies. In order for the advancement of reality service and industrial in our dear country, that a lot of scientific research touched upon the soil classification in the agricultural, commercial and other fields. No source and research can be found that touched upon the classification of land for industrial purposes directly. In this research specialized programs have been used such as geographic information system software The geographical information system permits the study of local distribution of phenomena, activities and the aims that can be determined in the loca
AS Salman, SK Hameed…, Karbala Journal of Physical Education Sciences, 2020
Abstract:
Through the research, I have briefly discussed the life of Imam Ali and
indicated the interpretation of the Qur'an by language and terminology, the
importance and characteristics of Sahabas', may Allah be pleased with them,
interpretation of the Qur'an. Also, the sayings of Imam Ali, wich included the
explanatory interpretation of the Qur'an by the Qura'n, Sunnah, language, and
correct opinion. I have also indicated the reasons of the Quran's revelation
and the readings of the Qura'n that had been received from him including the
interpretive meaning these readings revealed. Finally, I concluded my
research with the most important results and recommends.
إن النجاح في أداء المتطلبات الفنية والخططية في أي من الألعاب ألرياضيه يستوجب امتلاك العناصر الاساسيه المتعلقة بطبيعة الاداء ونوع الفعالية الرياضية الممارسة , لذا فان اغلب الألعاب الرياضية تعتمد على مكونات ألقدره التوافقيه والادراكيه الحسيه بوصفها احد العناصر الاساسيه في المستويات العليا لما توفره من قاعدة اقتران للصفات البدنيه والحر كيه وقدرات أجهزة الجسم الوظيفية , وفقا للأسس المعتمدة في بناء مهاراته, وع
... Show MoreUnderstanding the compatibility between spider silk and conducting materials is essential to advance the use of spider silk in electronic applications. Spider silk is tough, but becomes soft when exposed to water. Here we report a strong affinity of amine-functionalised multi-walled carbon nanotubes for spider silk, with coating assisted by a water and mechanical shear method. The nanotubes adhere uniformly and bond to the silk fibre surface to produce tough, custom-shaped, flexible and electrically conducting fibres after drying and contraction. The conductivity of coated silk fibres is reversibly sensitive to strain and humidity, leading to proof-of-concept sensor and actuator demonstrations.
This study sought to determine malformation caused by Ochratoxin-A (OTA) on mouse embryos. Twenty adult female white Swiss mice (mus msculus) were divided into four groups, with five females per group, and with one male placed with two females in a cage. Avaginal plug was observed in the early morning and the day of mating was considered as day of pregnancy followed by the first day of pregnancy. Three sub lethal concentrations of OTA were applied to the respective groups (other than the control), 1mg/kg, 2mg/kg and 4mg/kg. The animals were given 0.1 ml per 10 gm body weight per concentration of OTA once a day during days 7-14 of pregnancy. The control group animals were given distilled water. The pregnant mice were dissected, and the embry
... Show More