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.
Biodiesel production from microalgae depends on the biomass and lipid production. Both biomass and lipid accumulation is controlled by several factors. The effect of various culture media (BG11, BBM, and Urea), nutrients stress [nitrogen (N), phosphorous (P), magnesium (Mg) and carbonate (CO3)] and gamma (γ) radiation on the growth and lipid accumulation of Dictyochloropsis splendida were investigated. The highest biomass and lipid yield of D. splendida were achieved on BG11 medium. Cultivation of D. splendida in a medium containing 3000 mg L−1 N, or 160 mg L−1 P, or 113 mg L−1 Mg, or 20 mg L-1 CO3, led to enhanced growth rate. While u
... Show MoreThis study involved the effect of anew nickel (II) complexs with formla [NiL2(H2O)2].2.5ETOH where L=Bis[5-(p-nitrophenyL)-4-phenyL-1,2,4-traizole-3-dithocarbamato hydrazide] diaqua. nickel(II). Ethanol(2.5).and anti-cancer drug cyclophosphamide on specific actifity of two Liver enzymes (GOT,GPT) in the (Liver,kidney) tissues and on the creatinine Level in the kidney byUtilizing an invivosystem in femalmice.The result showed that inhibition in the activity of GPT and GOT enzymes in theLiver and in both nickel (II) complex and cyclophosphamide drug (CP) . mice weretreated with three doses (90,180,320) µg/mouse for three days for each group.The Liver show's the highest rate of GPT inhibition was about 97.43% at180µg/mouse regarding the ki
... Show MoreObjective: Atorvastatin therapy is now recommended for reduction of cardiovascular risk in type 2 diabetic patients (T2DM), based on convincing evidence of reductions in mortality and vascular events in major clinical outcome trials. The aim is to evaluate the effects of atorvastatin on proinflammatory markers (TNF-α, IL-6), HbA1c andleptin in obese patients with type 2 diabetes. Methods: Sixty fivenewly diagnosed T2DM patients were randomly allocated into 2 groups; group I treated with metformin only; in group II atorvastatin was added with metformin. Twenty healthy subjects were enrolled as control group. While maintaining their usual eating habits, fasting blood samples were collected at baseline and after 12 weeks of treatment. Results
... Show MoreSeveral adipokines are produced and secreted from adipose tissue, such as retinol binding protein-4, which triggers metabolic syndromes and insulin resistance. Retinol binding protein-4 transfers vitamin A or retinol in the blood. Higher levels of retinol binding protein-4 are interrelated with progress of metabolic disease, comprising obesity, metabolic syndrome, and type 2 diabetes mellitus. The present study investigates the role of retinol-binding protein-4 levels in type 2 diabetic Iraqi patients with metabolic syndrome. Sixty type 2 diabetic patients aged 40–53 years were examined. Of these 30 patients has metabolic syndrome and 30 without metabolic syndrome. The patients sampled were from the National Diabetes Center/ Mustansiriyah
... Show MoreBackground: Hypothyroidism is the most abundant thyroid disorder worldwide. For decades, levothyroxine was the main effective pharmacological treatment for hypothyroidism. A variety of factors can influence levothyroxine dose, such as genetic variations. Studying the impact of genetic polymorphisms on the administration of medications was risen remarkably. Different genetic variations were investigated that might affect levothyroxine dose requirements, especially the deiodinase enzymes. Deiodinase type 2 genetic polymorphisms’ impact on levothyroxine dose was studied in different populations. Objective: To examine the association of the two single nucleotide polymorphism (SNP)s of deiodinase type 2 (rs225013 and rs225014) and le
... Show MoreConsequence of thermal and concentration convection on peristaltic pumping of hyperbolic tangent nanofluid in a non‐uniform channel and induced magnetic field is discussed in this article. The brief mathematical modeling, along with induced magnetic field, of hyperbolic tangent nanofluid is given. The governing equations are reduced to dimensionless form by using appropriate transformations. Exact solutions are calculated for temperature, nanoparticle volume fraction, and concentration. Numerical technique is manipulated to solve the highly non‐linear differential equations. The roll of different variables is graphically analyzed in terms of concentration, temperature, volume fraction of nanoparticles, axial induced magnetic fie
... Show MoreThis study deals with free convection heat transfer for the outer surface of two
cylinders of the shape of (Triangular & Rectangular fined cylinders with 8-fins),
putted into two different spaces; small one with dimension of (Length=1.2m,
height=1m, width=0.9m) and large one with dimension of (Length=3.6m, height =3m,
width=2.7m). The experimental work was conducted with air as a heat transport
medium. These cylinders were fixed at different slope angles (0o, 30o, 60o and 90o)
.The heat fluxes were (279, 1012, 1958, 3005, 4419) W/m2, where heat transferred by
convection and radiation. In large space, the results show that the heat transfer from
the triangular finned cylinder is maximum at a slope angle equals
Roller-Compacted Concrete is a no-slump concrete, with no reinforcing steel, no forms, no finishing and wet enough to support compaction by vibratory rollers. Due to the effect of curing on properties and durability of concrete, the main purpose of this research is to study the effect of various curing methods (air curing, 7 days water curing, and permanent water curing) and porcelanite (local material used as an Internal Curing agent) with different replacement percentages of fine aggregate (volumetric replacement) on some properties of Roller-Compacted Concrete and to explore the possibility of introducing practical Roller-Compacted Concrete for road pavement with minimum requirement of curing. Specimens were sawed fro
... Show More