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 study seeks to use one of the techniques (Data mining) a (Logic regression) on the inherited risk through the use of style financial ratios technical analysis and then apply for financial fraud indicators,Since higher scandals exposed companies and the failure of the audit process has shocked the community and affected the integrity of the auditor and the reason is financial fraud practiced by the companies and not to the discovery of the fraud by the auditor, and this fraud involves intentional act aimed to achieve personal and harm the interests of to others, and doing (administration, staff) we can say that all frauds carried out through the presence of the motives and factors that help th
... Show MoreOne wide-ranging category of open source data is that referring to geospatial information web sites. Despite the advantages of such open source data, including ease of access and cost free data, there is a potential issue of its quality. This article tests the horizontal positional accuracy and possible integration of four web-derived geospatial datasets: OpenStreetMap (OSM), Google Map, Google Earth and Wikimapia. The evaluation was achieved by combining the tested information with reference field survey data for fifty road intersections in Baghdad, Iraq. The results indicate that the free geospatial data can be used to enhance authoritative maps especially small scale maps.
In this study the faunistic of lady beetles (Coleoptera, Coccinellidae) was studied in Mehriz region (Yazd province) during 2009-2010. The total number of specimens of coccinellid beetles were collected from 6 different localities having altitudes from 1420-2420 m. Altogether 11 species from 8 genera, 3 tribes and 3 subfamilies were collected and identified. External characters plus characteristics male and female genitalia were used in order to diagnose species. Seven species were recorded for the first time from Yazd province (marked*). Many species were predacious, preying on various species of aphids, mites and coccids. Some species were also sent to Dr. Helmut Fursch in Germany for identification or confirmation. T
... Show MoreThe aim of this study was to get monosodium glutamate (MSG) flavor, which was obtained from glutamic acid, that produced from local isolated from Bacillus subtilis EN3A1-P19U7 which genetically improved, from Bacillus subtilis EN3A1-P19U7, and applied in sausage chicken meat, mayonnaise and vegetable and lentil soup, it has been added MSG product in this study at different concentrations with the use of chicken broth cubes (Maggi) as a commercial flavor for comparison, and it was conducted sensory evaluation of these products and found that the addition of MSG product this study at the level of 0.6% to the sausage chicken and 0.6% to the mayonnaise and 0.15% to the vegetable and lentil soup, the results of sensory evaluation show not signif
... Show MoreBackground. “Polyetheretherketone (PEEK)” is a biocompatible, high-strength polymer that is well-suited for use in dental applications due to its unique properties. However, achieving good adhesion between PEEK and hydrophilic materials such as dental adhesives or cement can be challenging. Also, this hydrophobicity may affect the use of PEEK as an implant material. Surface treatment or conditioning is often necessary to improve surface properties. The piranha solution is the treatment of choice to be explored for this purpose. Methods. PEEK disks of 10 mm diameter and 2 mm thickness were used in this study. Those samples were divided into five groups (each group has five samples). The first is the control group, in which no
... Show MoreDue to the great evolution in digital commercial cameras, several studies have addressed the using of such cameras in different civil and close-range applications such as 3D models generation. However, previous studies have not discussed a precise relationship between a camera resolution and the accuracy of the models generated based on images of this camera. Therefore the current study aims to evaluate the accuracy of the derived 3D buildings models captured by different resolution cameras. The digital photogrammetric methods were devoted to derive 3D models using the data of various resolution cameras and analyze their accuracies. This investigation involves selecting three different resolution cameras (low, medium and
... Show MoreRecently, several concepts and expressions have emerged that have often preoccupied the world . around the concept of environment and sustainability. This is due to the negative and irresponsible impact of man and his innovations in various industrial and technological fieldsthat have damaged the natural environment. Architecture and cities at the broader level are some of the man made components that caused these negative impacts and in the same time affected by them. What distinguishes architectural and urban projects is the consumption of large . quantities of natural resources and production larger amounts of waste and pollution, along the life of these projects. At the end of the twentieth century and the beginning of the twenty-fir
... Show MoreThis study suggests using the recycled plastic waste to prepare the polymer matrix composite (PMCs) to use in different applications. Composite materials were prepared by mixing the polyester resin (UP) with plastic waste, two types of plastic waste were used in this work included polyethylene-terephthalate (PET) and Polyvinyl chloride (PVC) with varies weight fractions (0, 5, 10, 15, 20 and 25 %) added as a filler in flakes form. Charpy impact test was performed on the prepared samples to calculate the values of impact strength (I.S). Flexural and hardness tests were carried out to calculate the values of flexural strength and hardness. Acoustic insulation and optical microscope tests were carried out. In general, it is found that UP/PV
... Show More