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.
This paper is concerned with introducing an explicit expression for orthogonal Boubaker polynomial functions with some important properties. Taking advantage of the interesting properties of Boubaker polynomials, the definition of Boubaker wavelets on interval [0,1) is achieved. These basic functions are orthonormal and have compact support. Wavelets have many advantages and applications in the theoretical and applied fields, and they are applied with the orthogonal polynomials to propose a new method for treating several problems in sciences, and engineering that is wavelet method, which is computationally more attractive in the various fields. A novel property of Boubaker wavelet function derivative in terms of Boubaker wavelet themsel
... Show MoreThe aim of this study was to identify the effectiveness of using generative learning model in learning kinetic series on rings and horizontal bar in artistic gymnastics for men ,Also, the two groups were better in learning the two series of movements on the rings and horizontal bar . The experimental method was used to design two parallel groups with pretested and posttest .The sample included third graders at the College of Physical Education and Sports Sciences - University of Baghdad ,The third class (d) was chosen to represent the control group that applied the curriculum in the college, with (12) students per group. After conducting the tribal tests, the main experiment was carried out for (8) weeks at the rate of two units per week di
... Show MoreThis study was set out to investigate factors affecting labor productivity on construction in the north of Iraq (Kurdistan) and to rank all the factors based on engineers, contractors, and designer’s opinions. 76 factors were analyzed based on previous literature and a pilot study. Next, by using online Google Form, a questionnaire form was created and sent to people who have experience in the construction industry. Afterward, the questionnaire form was sent to targeted people by email and social media apps. Factors were divided into nine groups “Management, Technical and Technology, Human and Workforce, Leadership, Motivation, Safety, Time, Material and Equipment, and External”. However, 202 respondents participated in this study,
... Show MoreAt the level of both individuals and companies, Wireless Sensor Networks (WSNs) get a wide range of applications and uses. Sensors are used in a wide range of industries, including agriculture, transportation, health, and many more. Many technologies, such as wireless communication protocols, the Internet of Things, cloud computing, mobile computing, and other emerging technologies, are connected to the usage of sensors. In many circumstances, this contact necessitates the transmission of crucial data, necessitating the need to protect that data from potential threats. However, as the WSN components often have constrained computation and power capabilities, protecting the communication in WSNs comes at a significant performance pena
... Show MoreThe involvement of maxillofacial tissues in SARS‐CoV‐2 infections ranges from mild dysgeusia to life‐threatening tissue necrosis, as seen in SARS‐CoV‐2‐associated mucormycosis. Angiotensin‐converting enzyme 2 (ACE2) which functions as a receptor for SARS‐CoV‐2 was reported in the epithelial surfaces of the oral and nasal cavities; however, a complete understanding of the expression patterns in deep oral and maxillofacial tissues is still lacking.
The immunohistochemical expression of ACE2 was analyzed in 95 specimens from maxillofacial tissues and 10 specimens o
In this paper, the theoretical cross section in pre-equilibrium nuclear reaction has been studied for the reaction at energy 22.4 MeV. Ericson’s formula of partial level density PLD and their corrections (William’s correction and spin correction) have been substituted in the theoretical cross section and compared with the experimental data for nucleus. It has been found that the theoretical cross section with one-component PLD from Ericson’s formula when doesn’t agree with the experimental value and when . There is little agreement only at the high value of energy range with the experimental cross section. The theoretical cross section that depends on the one-component William's formula and on-component corrected to spi
... Show MoreThe aim of this research to study.
The dimensions of organizational learning have been defined(learning dynamics, individuals empowerment, knowledge management and technology application) as well as the dimensions of learning organization have been defined (culture values, knowledge transfer, communication and employee characteristics), Asset completion questionnaire was used to collect data of this research from a purposely sample represent forty employees who works in Iraqi Planning Ministry at different positions. The research divided to four parts :
The first to the research methodology, the second to the theoretical review o
... Show MorePurpose – The main purpose of this research is to highlight the main role of strategic leadership skills for top managements in accessing to effective management in accordance with the (VUCA Prime) methodology in (VUCA) environment as Miniature virtual environment, which refers to (Volatility), (Uncertainty), (Complexity), and (Ambiguity).
methodology – To achieve the research objective, this study selected the quantitative approach in research design, Questionnaire was used as the main instrument for data collection, the sample comprised the opinion poll (106) individual who functions as a head department. (Structural equation modelling by (Smart Pls3)
... Show More