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.
Strategic Cost Management Tools Under Technological Development and Change in Customer Tastes Critical Studies
The huge amount of documents in the internet led to the rapid need of text classification (TC). TC is used to organize these text documents. In this research paper, a new model is based on Extreme Machine learning (EML) is used. The proposed model consists of many phases including: preprocessing, feature extraction, Multiple Linear Regression (MLR) and ELM. The basic idea of the proposed model is built upon the calculation of feature weights by using MLR. These feature weights with the extracted features introduced as an input to the ELM that produced weighted Extreme Learning Machine (WELM). The results showed a great competence of the proposed WELM compared to the ELM.
Abstract
The main problem of the study lies in the lack of a clear perception among the study sample about the impact of digital marketing tools on legal liquidity. Legal) of the International Development Bank for Investment and Finance and to achieve the objectives of the research, the method of observation and survey was used in measuring the dimensions of digital marketing. As for banking liquidity, the reports and financial statements of the bank were used as the research sample, as well as the use of the statistical analysis program SPSS in the statement of the relationship The study concluded, in summary, the following: Mar
... Show MoreMany 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), Decision Tre
... Show MoreMany 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 MoreA Laced Reinforced Concrete (LRC) structural element comprises continuously inclined shear reinforcement in the form of lacing that connects the longitudinal reinforcements on both faces of the structural element. This study conducted a theoretical investigation of LRC deep beams to predict their behavior after exposure to fire and high temperatures. Four simply supported reinforced concrete beams of 1500 mm, 200 mm, and 240 mm length, width, and depth, respectively, were considered. The specimens were identical in terms of compressive strength ( 40 MPa) and steel reinforcement details. The same laced steel reinforcement ratio of 0.0035 was used. Three specimens were burned at variable durations and steady-state temperatures (one
... Show MoreCombining different treatment strategies successively or simultaneously has become recommended to achieve high purification standards for the treated discharged water. The current work focused on combining electrocoagulation, ion-exchange, and ultrasonication treatment approaches for the simultaneous removal of copper, nickel, and zinc ions from water. The removal of the three studied ions was significantly enhanced by increasing the power density (4–10 mA/cm2) and NaCl salt concentration (0.5–1.5 g/L) at a natural solution pH. The simultaneous removal of these metal ions at 4 mA/cm2 and 1 g NaCl/L was highly improved by introducing 1 g/L of mordenite zeolite as an ion-exchanger. A remarkable removal of heavy metals was reported
... Show MoreCombining different treatment strategies successively or simultaneously has become recommended to achieve high purification standards for the treated discharged water. The current work focused on combining electrocoagulation, ion-exchange, and ultrasonication treatment approaches for the simultaneous removal of copper, nickel, and zinc ions from water. The removal of the three studied ions was significantly enhanced by increasing the power density (4–10 mA/cm2) and NaCl salt concentration (0.5–1.5 g/L) at a natural solution pH. The simultaneous removal of these metal ions at 4 mA/cm2 and 1 g NaCl/L was highly improved by introducing 1 g/L of mordenite zeolite as an ion-exchanger. A remarkable removal of heavy metals was reported
... Show More