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.
Many academics have concentrated on applying machine learning to retrieve information from databases to enable researchers to perform better. A difficult issue in prediction models is the selection of practical strategies that yield satisfactory forecast accuracy. Traditional software testing techniques have been extended to testing machine learning systems; however, they are insufficient for the latter because of the diversity of problems that machine learning systems create. Hence, the proposed methodologies were used to predict flight prices. A variety of artificial intelligence algorithms are used to attain the required, such as Bayesian modeling techniques such as Stochastic Gradient Descent (SGD), Adaptive boosting (ADA), Deci
... Show MoreBackground/Objectives: The purpose of this study was to classify Alzheimer’s disease (AD) patients from Normal Control (NC) patients using Magnetic Resonance Imaging (MRI). Methods/Statistical analysis: The performance evolution is carried out for 346 MR images from Alzheimer's Neuroimaging Initiative (ADNI) dataset. The classifier Deep Belief Network (DBN) is used for the function of classification. The network is trained using a sample training set, and the weights produced are then used to check the system's recognition capability. Findings: As a result, this paper presented a novel method of automated classification system for AD determination. The suggested method offers good performance of the experiments carried out show that the
... Show MoreThe administrative leadership relies on a variety of behavioral paths in the functional areas in which it operates, and thus indicates its ability and thus has the upper hand in the organizational events, and in such a way that it can draw lessons and evaluate the results so as to test the expectations within the framework of the changes.
Does the study sample leadership have the characteristics that enable it to contain both crises and environmental stresses?
The aim of the study was to determine the location of the administrative leaders in the system of the study sample from the issue of crises and environmental pressures. The study concluded with a number of conclusions, the m
... Show MoreThe research aims to shed light on the concept of Visibal management and its reflection on the organizational culture of the organization. The visual administration is a modern administrative method that contributes to the renewal and development of the organization's reality through surveying the opinions of a sample of 61 employees in the R & D / Ministry of Higher Education and Scientific Research. (130) individuals. The questionnaire was used as a main tool for collecting data and information, and their answers were analyzed using the SPSS program in data entry and analysis. The most important tools are computational circles, standard deviations, method of analysis and regression equation. There is a possibility to apply
... Show MoreThis study deals with 3rd-order difference equations with alternating coefficients (NDEAC) with a focus on studying the oscillatory properties of their solutions. An additional sufficient condition can be derived to ensure continuity for the oscillatory solution in equations. An applied example is presented to illustrate the theoretical result obtained.
Dyes are extensively water-soluble and toxic chemicals. The disposing of wastewater rich with such chemicals has severely impacted surface water quality (rivers and lakes). In the current study, an anionic dye, methyl orange, were extracted from wastewater fluids using bulk liquid membranes supplemented with an anionic carrier (Aliquat 336 (QCI)). Parameters including solvent type (carbon tetrachloride and chloroform), membrane stirring speed (100-250 rpm), mixing speed of both phases (50-100 rpm), The feed pH (2-12) and implemented temperature (35-60 °C) were thoroughly analyzed to determine the effect of such variables on extraction effectiveness. Furthermore, the effect of methyl orange (10-50 ppm) in the feed stage and NaOH (0
... Show MoreThis study aims to identify the teaching problems that teachers of students with intellectual disabilities face, in addition to exploring the solutions suggested by them in order to overcome such problems or challenges. The researchers used a qualitative approach in order to understand the teachers' perceptions about these problems in a more in-depth way. The interview tools (in-depth and semi-structured interviews) were used to collect data from (3) female teachers from special education programs in the Asir region. The results revealed a number of themes including problems related to students, teachers and the teaching methods they use, curricula, school environment, and school administration. Moreover, the results indicated that famil
... Show MoreA Survey Study Of A Sample Of The Public Of Baghdad Governorate
The current study aimed to identify the most prominent psychological and behavioral repercussions of the exposure of the elderly to the news of the Corona pandemic and to determine the mechanisms of their exposure. On an intended sample on both sides of (Al-Karkh and Al-Rasafa) and the simple random sample was adopted to choose the place of distribution of the questionnaire and the intentional sample.
The research concluded several results, most TV news is still a primary source of information and that most of the sample move between stations to see more information about the pandemic and that the presentation of views confuses the elderly ،There
... Show More