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.
Recently, the application of geosynthetics in the reinforcement of weak subgrade is expanded dramatically. However, selection of the geo-material that fits site conditions and soil type is crucial to achieving the success of the overall performance of such improvement. Also, the road life and cost construction are significant keys for evaluating this type of ground treatment. This paper presents an overview of the subgrade strengthening with geosynthetics to acquire a better understanding of the technique and to provide a clear guide for transportation and geotechnical engineers. The rutting failure along with its main causes are highlighted briefly. The types of geosynthetics, their applications and
Statement of the Problem. The use of orthodontic fixed appliances may adversely affect oral health leading to demineralizing lesions and the development of gingival problems. Aims of the Study. The study aimed to coat orthodontic archwires with chlorhexidine hexametaphosphate nanoparticles (CHX-HMP NPs) and to evaluate the elusion of CHX from CHX-HMP NPs. Materials and Methods. A solution of CHX-HMP nanoparticles with an overall concentration of 5 mM for both CHX and HMP was prepared, characterized (using atomic force microscope and Fourier transformation infrared spectroscopy), and used to coat orthodontic stainless steel (SSW) and NiTi archwires (NiTiW). The coated segments were characterized (using scanning electron microscopy
... Show MoreIn this paper, we introduce an exponential of an operator defined on a Hilbert space H, and we study its properties and find some of properties of T inherited to exponential operator, so we study the spectrum of exponential operator e^T according to the operator T.
Multiple eliminations (de-multiple) are one of seismic processing steps to remove their effects and delineate the correct primary refractors. Using normal move out to flatten primaries is the way to eliminate multiples through transforming these data to frequency-wavenumber domain. The flatten primaries are aligned with zero axis of the frequency-wavenumber domain and any other reflection types (multiples and random noise) are distributed elsewhere. Dip-filter is applied to pass the aligned data and reject others will separate primaries from multiple after transforming the data back from frequency-wavenumber domain to time-distance domain. For that, a suggested name for this technique as normal move out- frequency-wavenumber domain
... Show MoreIn this research want to make analysis for some indicators and it's classifications that related with the teaching process and the scientific level for graduate studies in the university by using analysis of variance for ranked data for repeated measurements instead of the ordinary analysis of variance . We reach many conclusions for the
important classifications for each indicator that has affected on the teaching process. &nb
... Show MoreThis study has been accomplished by testing three different models to determine rocks type, pore throat radius, and flow units for Mishrif Formation in West Qurna oilfield in Southern Iraq based on Mishrif full diameter cores from 20 wells. The three models that were used in this study were Lucia rocks type classification, Winland plot was utilized to determine the pore throat radius depending on the mercury injection test (r35), and (FZI) concepts to identify flow units which enabled us to recognize the differences between Mishrif units in these three categories. The study of pore characteristics is very significant in reservoir evaluation. It controls the storage mechanism and reservoir fluid prope
The current research aimed to conducting two experiments to study the effect of coating hatching eggs with nano-titanium dioxide (nano-TiO2) and nano-silica dioxide (nano-SiO2) particles and their mixture with carboxymethyl cellulose (CMC) on the characteristics of hatching percentage, embryo growth inside the egg. The study was conducted in the Department of Animal Production, College of Agriculture, Tikrit University for the period from 19/3/2023 to 17/9/2024. It aimed to evaluate the coating of hatching eggs with Nano-TiO2 and Nano-SiO2 particles and their mixture with carboxymethyl cellulose CMC on the qualities of hatching percentage, embryo growth inside the egg, as well as trying to obtain the best and longest storage method for fert
... Show MoreThis experiment was carried out at a private field in the eastern Radwaniyah Baghdad for the fall season 2020/2021 and spring 2021 to study the effects of adding mineral fertilizers, spraying salicylic acid and amino acids on some growth traits and yield of industrial potato plants. 200 kg N h-1 , 100 kg P2O5 h-1, 100 kg K2O h-1 and F2 consist of 275 kg N h-1, 180 kg P2O5 h-1, 200 K2O h-1 and F3 consist of 350 kg N h-1, 360 kg P2O5 h-1, 300 K2O h-1 and salicylic acid in three concentrations of 0,50 and 100 mg L-1 ( S1, S2, S3) and amino acids in three concentrations of 0, 1.25 and 2.5 ml L-1 ( A1, A2 , A3) It was carried out as a factorial split plot experiment, where the fertilizer levels (F1, F2 and F3) are in the main plot and th
... Show More