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 rapid development of Internet of Things (IoT) devices and their increasing numbers have caused a tremendous increase in network traffic and a wider range of cyber-attacks. This growing trend has complicated the detection process for traditional intrusion detection systems and heightened the challenges faced by these devices, such as imbalanced and large training data. This study presents a cohesive methodology of a series of intelligent techniques to prepare clean and balanced data for training the first (core) layer of a robust hierarchical intrusion detection system. The methodology was built by cleaning and compressing the data using an Autoencoder and preparing a strong latent space for balancing using a hybrid method that combines
... Show MoreThis research describes a new model inspired by Mobilenetv2 that was trained on a very diverse dataset. The goal is to enable fire detection in open areas to replace physical sensor-based fire detectors and reduce false alarms of fires, to achieve the lowest losses in open areas via deep learning. A diverse fire dataset was created that combines images and videos from several sources. In addition, another self-made data set was taken from the farms of the holy shrine of Al-Hussainiya in the city of Karbala. After that, the model was trained with the collected dataset. The test accuracy of the fire dataset that was trained with the new model reached 98.87%.
The current research focuses on the extent to which the strategic orientation(entrepreneurial orientation, customer orientation, technology orientation, learning orientation, and investment orientation) affects the learning organization (building common vision, systemic thinking, personal dominance, mental models, team learning)The first hypothesis to test the connection relation between research variables and The second hypothesis was to test the relationship between these variables. In order to ascertain the validity of the hypotheses, the research was based on a questionnaire questionnaire prepared according to a number of In addition to building a fifth sub-variable for the strategic orientation (investment orientation) based
... Show MoreWith the development of communication technologies, the use of wireless systems in biomedical implanted devices has become very useful. Bio-implantable devices are electronic devices which are used for treatment and monitoring brain implants, pacemakers, cochlear implants, retinal implants and so on. The inductive coupling link is used to transmit power and data between the primary and secondary sides of the biomedical implanted system, in which efficient power amplifier is very much needed to ensure the best data transmission rates and low power losses. However, the efficiency of the implanted devices depends on the circuit design, controller, load variation, changes of radio frequency coil’s mutual displacement and coupling coef
... Show MoreDue to the remarkable progress in photovoltaic technology, enhancing efficiency and minimized the costs have emerged as global challenges for the solar industry. A crucial aspect of this advancement involves the creation of solar cell antireflection coating, which play a significant role in minimizing sunlight reflection on the cell surface. In this study, we report on the optimization of the characteristics of CeO2 films prepared by pulsed laser deposition through the variation of laser energy density. The deposited CeO2 nanostructure films have been used as an effective antireflection coating (ARC) and light-trapping morphology to improve the efficiency of silicon crystalline solar cell. The film’s thickness increases as laser fluence i
... Show MoreSystems on Chips (SoCs) architecture complexity is result of integrating a large numbers of cores in a single chip. The approaches should address the systems particular challenges such as reliability, performance, and power constraints. Monitoring became a necessary part for testing, debugging and performance evaluations of SoCs at run time, as On-chip monitoring is employed to provide environmental information, such as temperature, voltage, and error data. Real-time system validation is done by exploiting the monitoring to determine the proper operation of a system within the designed parameters. The paper explains the common monitoring operations in SoCs, showing the functionality of thermal, voltage and soft error monitors. The different
... Show MoreTraffic classification is referred to as the task of categorizing traffic flows into application-aware classes such as chats, streaming, VoIP, etc. Most systems of network traffic identification are based on features. These features may be static signatures, port numbers, statistical characteristics, and so on. Current methods of data flow classification are effective, they still lack new inventive approaches to meet the needs of vital points such as real-time traffic classification, low power consumption, ), Central Processing Unit (CPU) utilization, etc. Our novel Fast Deep Packet Header Inspection (FDPHI) traffic classification proposal employs 1 Dimension Convolution Neural Network (1D-CNN) to automatically learn more representational c
... Show MoreIf the State attaches great importance to its foreign relations and intends to strengthen them in order to ensure the achievement of the highest national goals and interests. External relations between countries are one of the most prominent features of foreign policy, which depends on a combination of internal and external factors, the modern relations between Tunisia and Tunisia, which goes back to the pre-independence of Tunisia, when Iraq was a supporter of Tunisia's independence from France in the 1940s, Although these relations did not cause any disturbance by the two countries, but they remained weak relations did not develop in all areas except the sports and cultural field, which we will determine the reasons and the pos
... Show More