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.
The aim of this book is to present a method for solving high order ordinary differential equations with two point boundary condition of the different kind, we propose semi-analytic technique using two-point osculatory interpolation to construct polynomial solution. The original problem is concerned using two-points osculatory interpolation with the fit equal numbers of derivatives at the end points of an interval [0 , 1] . Also, we discussion the existence and uniqueness of solutions and many examples are presented to demonstrate the applicability, accuracy and efficiency of the methods by compared with conventional method .i.e. VIDM , Septic B-Spline , , NIM , HPM, Haar wavelets on one hand and to confirm the order convergence on the other
... Show MoreThe musical templates are the fundamental reason for admiration and interest among a lot of cultural and societal medias because of the beauty of it's melodic value, where a lot of Iraqi singing and music specialists and composers try to consolidate daily life idea and translate it into music in a manner that preserve the template and rhythm horizontally and vertically, through the involvement of the scientific and philosophical concepts and theories of modern thought to reach the recipient,, as is the theory of form (Gestalt), one of the most important theories that stretched to their interpretations to the field in general art and in music science, especially, as an area that can be manifested as partial components template music, conn
... Show MoreIn probability theory generalizing distribution is an important area. Several distributions are inappropriate for data modeling, either symmetrical, semi-symmetrical, or heavily skewed. In this paper, a new compound distribution with four parameters called Marshall Olkin Marshall Olkin Weibull (MOMOWe) is introduced. Several important statistical properties of new distribution were studied and examined. The estimation of unknown four parameters was carried out according to the maximum likelihood estimation method. The flexibility of MOMOWe distribution is demonstrated by the adoption of two real datasets (semi-symmetric and right-skewed) with different information fitting criteria. Su
Sorting and grading agricultural crops using manual sorting is a cumbersome and arduous process, in addition to the high costs and increased labor, as well as the low quality of sorting and grading compared to automatic sorting. the importance of deep learning, which includes the artificial neural network in prediction, also shows the importance of automated sorting in terms of efficiency, quality, and accuracy of sorting and grading. artificial neural network in predicting values and choosing what is good and suitable for agricultural crops, especially local lemons.
Micro metal forming has an application potential in different industrial fields. Flexible tool-assisted sheet metal forming at micro scale is among the forming techniques that have increasingly attracted wide attention of researchers. This forming process is a suitable technique for producing micro components because of its inexpensive process, high quality products and relatively high production rate. This study presents a novel micro deep drawing technique through using floating ring as an assistant die with flexible pad as a main die. The floating ring designed with specified geometry is located between the process workpiece and the rubber pad. The function of the floating ring in this work is to produce SS304 micro cups with profile
... Show MoreSoftware-defined networking (SDN) is an innovative network paradigm, offering substantial control of network operation through a network’s architecture. SDN is an ideal platform for implementing projects involving distributed applications, security solutions, and decentralized network administration in a multitenant data center environment due to its programmability. As its usage rapidly expands, network security threats are becoming more frequent, leading SDN security to be of significant concern. Machine-learning (ML) techniques for intrusion detection of DDoS attacks in SDN networks utilize standard datasets and fail to cover all classification aspects, resulting in under-coverage of attack diversity. This paper proposes a hybr
... Show More