Search

Code smells violate software development principles that make the software more prone to errors and changes. Researchers have developed code smell detectors using manual and semi-automatic methods to identify these issues. However, three key challenges have limited the practical use of these detectors: developers' subjective perceptions of code smells, lack of consensus in the detection process, and difficulty in setting appropriate detection thresholds. While code smell detection using machine learning has progressed significantly, there still appears to be a gap in understanding the effective utilization of deep learning (DL) approaches. This paper aims to review and identify current methods for code smell detection using DL techniques. A systematic literature review is conducted on 35 primary studies from a collection of 8739 publications between 2013 and the present. The analysis reveals that common code smells detected include Feature Envy, God Classes, Long Methods, Complex Classes, and Large Classes. The most popular DL algorithms used are Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN), often combined with other techniques for better results. Algorithms that train models on large datasets with fewer independent variables demonstrate exemplary performance. The paper also highlights open issues and provides guidelines for future metric identification and selection research. [ABSTRACT FROM AUTHOR]

Massive amounts of real-time streaming network data are generated quickly because of the exponential growth of applications. Analyzing patterns in generated flow traffic streaming offers benefits in reducing traffic congestion, enhancing network management, and improving the quality of service management. Processing massive volumes of generated traffic poses more challenges when data traffic encryption is raised. Classifying encrypted network traffic in real-time with deep learning networks has received attention because of their excellent performance. The substantial volume of incoming packets, characterized by high speed and wide variety, puts real-time traffic classification within the domain of big data problems. Classifying traffic with high speed and accuracy is a significant challenge in the era of big data. The real-time nature of traffic intensifies deep learning networks, necessitating a considerable number of parameters, layers, and resources for optimal network training. Until now, various datasets have been employed to evaluate the effectiveness of previous methods for classifying encrypted traffic. The primary objective has been to enhance accuracy, precision, and F1-measure. Presently, encrypted traffic classification performance depends on pre-existing datasets. The learning and testing phases are done offline, and more research is needed to investigate the feasibility of these methods in real-world scenarios. This paper examines the possibility of a tradeoff between evaluating the model's effectiveness, execution time, and utilization of processing resources when processing stream-based input data for traffic classification. We aim to explore the feasibility of establishing a tradeoff between these factors and determining optimal parameter settings. This paper used the ISCX VPN-Non VPN 2016 public dataset to evaluate the proposed method. All packets from the dataset were streamed continuously through Apache Kafka to the classification framework. Numerous experiments have been designed to demonstrate the efficacy of the proposed method. The experimental results show that the proposed method outperforms the baseline methods by 11% in the F1-measure when the number of workers is two and by 25% when the number of workers is equal to 32. [ABSTRACT FROM AUTHOR]

The dominant intrusion detection models in internet of things industrial internet of things cybersecurity use network-based datasets. The Modbus protocol is one of the most often targeted protocols and cyberattacks against IoT/IIoT devices have grown to be a major threat in recent years. Due to the intricacy of the protocol and the quick evolution of cyber threats, detecting these attacks using conventional techniques might be difficult. This paper proposes an architecture that consistently outperforms the state-of-the-art methods of performing intrusion Detection that includes binary classification of whether an intrusion occurred or not and multi-class classification that classifies the different types of attacks using an embedding layer in a neural network to model the register values. The best accuracy results were obtained with a convolutional neural network, with an accuracy of 98.91% in the Modbus Binary dataset, a fully connected neural network with an accuracy of 98.06% in the multi-class classification of the Modbus dataset, and long short-term memory neural networks with an accuracy of 99.97%, 99.7%, and 80.20% in Binary, multi-class, and multi-class sub-categories, respectively which conclude that the proposed architecture performs consistently better than the control NN. Three NN are designed with and without the proposed architecture. All experiments performed in this paper conclude that the proposed architecture performs consistently better than the control NN. This paper shows that a NN with an embedding function can effectively be used to model whether an attack occurred on a device and the class of attack that occurred. This network can be utilized in the future to lessen DoS attacks and other types of network attacks. The network will be able to protect itself against a lot of damage if attacks can be predicted either before they occur or at the same moment they are launched. [ABSTRACT FROM AUTHOR]

With the development of collaborative edge computing (CEC), the manufacturing market is gradually moving toward large-scale, multi-scenario, and dynamic directions. The existing scheduling strategies based on machine learning or deep learning are only applicable to specific scenarios, which is difficult to meet the requirements of dynamic real-time scheduling in multiple scenarios. The proposed digital twin technology provides a new solution for real-time scheduling of multiple scenarios. In this paper, a digital twin-oriented multi-scene real-time scheduler (GNN-RL) is proposed. This scheduler converts task sequences into node trees and sets up two learning layers. The first layer is an online learning representation layer, which uses GNN to learn node features of embedded structures in real time to boost large instances without additional training. The second layer is the online learning policy layer, which introduces imitation learning mappings into optimal scheduling behavior policies adapted to multiple scenarios. Finally, our approach is validated in several scenarios in 3D digital twin factories, such as computationally intensive, delay-sensitive, and task-urgent scenarios. Since the scheduler proposed in this paper learns general features of the embedding graph rather than instance-specific features, it has good generality and scalability, with good generalization and scalability, outperforming other scheduling rules and schedulers on various benchmarks. [ABSTRACT FROM AUTHOR]

Understanding users' exact feelings and enhancing enterprise customer relationship management depend heavily on emotion and sentiment analysis in intelligent customer service conversations. However, the research that is currently available analyzes either emotion or sentiment. This paper proposes a multi-task ensemble model that can perform multiple tasks of emotion and sentiment analysis simultaneously. This ensemble model via a multi-layer perceptron (MLP) network develops three deep learning models based on convolutional neural network (CNN), long short-term memory (LSTM), and gated recurrent unit (GRU) for predicting intelligent customer service dialogue emotion and sentiment analysis, including "emotion classification and intensity", "valence and arousal for emotion", "valence and arousal for sentiment", and "3-class categorical classification for sentiment". The underlying problems cover two granularity analysis (i.e., coarse-grained and fine-grained) in the intelligent customer service domain. Experimental results suggest that the proposed multi-task ensemble model outperforms the single-task framework, and the method performs well in emotion and sentiment analysis tasks in intelligent service conversation. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT) provides a collaborative infrastructure to communicate smart devices with cloud-edge healthcare applications, medical devices, wearable biosensors, etc. On the other hand, crowd counting as one of computer vision approaches is an emerging topic to detect any objects with static or dynamic mobility in the IoT environments. Smart crowd counting enables pattern recognition for many intelligent applications such as microbiology, surveillance, healthcare systems, crowdedness estimation, and other environmental case studies. According to complicated capturing systems in the IoT environments, crowd counting methods can influence on performance of object detection in the critical case studies using Artificial Intelligence (AI)-based approaches such as machine learning, deep learning, collaborative learning, fuzzy logic and meta-heuristic algorithms. This paper provides a new comprehensive technical analysis for existing AI-based crowd counting approaches in healthcare and medical systems, biotechnology and IoT environments. Meanwhile, it presents a discussion on the existing case studies with respect to analyzing technical aspects and applied algorithms to enhance pattern prediction factors. Finally, some new innovative efforts and challenges are presented for new research upcoming and open issues. [ABSTRACT FROM AUTHOR]

Given the rise in demand for cloud computing in the modern era, the effectiveness of resource utilization is eminent to decrease energy footprint and achieve economic services. With the emerging machine learning and artificial intelligence techniques to model and predict, it is essential to explore a principal method that provides the best solution for the accurate provisioning of forthcoming requests in a cloud data center. Recent studies used machine learning and other advanced analytics to predict resource usage; however, these do not consider long-range dependencies in the time series, which is essential to capture for better prediction. Further, they show limitations in handling noise, missing values, and outliers in datasets. In this paper, we explored the problem by studying three techniques that enabled us to answer improvements in short-term forecasting of physical machines' resource usage if the above factors are considered. We evaluated the predictions using Transformer and Informer deep learning models that cover the above aspects and compared them with the Long short-term memory (LSTM) model. We used a real-world Google cluster trace usage dataset and employed Balanced Iterative Reducing and Clustering using Hierarchies (BIRCH) algorithm to select heterogeneous machines. The evaluation of the three models depicts that the Transformer architecture that considers long-range dependencies in time series and shortcomings with datasets shows improvement in forecasting with 14.2% reduction in RMSE than LSTM. However, LSTM shows better results for some machines than the Transformer, which depicts the importance of input sequence order. The Informer model, which considers both dependencies and is a hybrid of LSTM and Transformer, outperformed both models with 21.7% from LSTM and 20.8% from Transformer reduction in RMSE. The results also depict Informer model consistently performs better than the other models across all subsets of the dataset. Our study proves that considering long-range dependencies and sequence ordering for resource usage time series improves the prediction. [ABSTRACT FROM AUTHOR]

Availability is one of the primary goals of smart networks, especially, if the network is under heavy video streaming traffic. In this paper, we propose a deep learning based methodology to enhance availability of video streaming systems by developing a prediction model for video streaming quality, required power consumption, and required bandwidth based on video codec parameters. The H.264/AVC codec, which is one of the most popular codecs used in video steaming and conferencing communications, is chosen as a case study in this paper. We model the predicted consumed power, the predicted perceived video quality, and the predicted required bandwidth for the video codec based on video resolution and quantization parameters. We train, validate, and test the developed models through extensive experiments using several video contents. Results show that an accurate model can be built for the needed purpose and the video streaming quality, required power consumption, and required bandwidth can be predicted accurately which can be utilized to enhance network availability in a cooperative environment. [ABSTRACT FROM AUTHOR]

The technology of big data analysis and artificial intelligence deep learning has been actively cross-combined with various fields to increase the effect of its original low single field. Precision components commonly used in electronic products use changes in the conductivity of semiconductors to process information. This study aims to review key milestones and recent developments in the semiconductor industry using artificial intelligence methods. For this systematic review, we searched academic networks between 2015 and 2022, including Nature, Elsevier, Springer, Taylor & Francis Online, Multidisciplinary Digital Publishing Institute, and the Institute of Electrical and Electronics Engineers. The literature reviewed is based on conference proceedings and journal articles, specifically covering the key achievements of the discussion paper, the key technologies used, experimental results, opportunities, and future research pathways. After searching on an academic website, we selected six major studies. In five of these studies, visual object detection, surface defect detection, machine production scheduling application, fault diagnosis and prediction, and monitoring of the manufacturing process were made using artificial neural networks, machine learning methods, and hybrid models. In addition, the studies covered independent, single methods or used more than two types of technologies for performance comparison. Finally, we reviewed the strengths and weaknesses of the literature. We also analysed various datasets, acquisition systems, and experimental scenarios. The review shows that as the number of studies conducted in manufacturing continues to increase, more research is needed to unearth key information that is often overlooked, all of which are challenges in refining science and overcoming real-world scenarios. [ABSTRACT FROM AUTHOR]

As the scale and complexity of networks grow rapidly, management, maintenance, and optimization of them are becoming increasingly challenging tasks for network administrators. Software-Defined Networking (SDN) was introduced to facilitate these tasks as it offers logically centralized control, a global view of the network, and software-based traffic analysis, thus, it is widely adopted of SDN to manage large-scale networks. On the other hand, SDN is not immune to cyber attacks. In fact, its centralized architecture makes it more vulnerable to certain types of attacks, such as denial of service. Various attack mitigation strategies are proposed to strengthen the security of SDNs including statistical, threshold-based, and Machine Learning (ML) methods. Among them, Deep Learning (DL)-based models attained the best results as they were able to extract the complex relationship between input parameters and output that could not be achieved with other solutions. Hence, this paper presents a comprehensive survey of the literature on the utilization of different DL algorithms for the security of SDN. We first explain the types of attacks that SDNs are exposed to, then present papers that applied DL to detect and/or mitigate these attacks. We further discuss the public datasets used to train DL models and evaluate their advantages and disadvantages. Finally, we share insights into future research directions to improve the efficiency of DL methods for SDN security. [ABSTRACT FROM AUTHOR]

Nowadays, Internet applications with different network resource requirements have been growing exponentially; therefore, the network equipment needs to be configured consistently based on application resource requirements. In this paper, Software Defined Network (SDN) is used to make the network more programmable, flexible and agile to develop the proposed model. Flow discrimination and optimized resource allocation are significant and challenging items in providing the required network resources for each flow. In this paper, we propose a model composed of (1) network flows' types discriminator and (2) the optimized resource allocation based on the flow classes. The applications are clustered into four groups according to the network resource requirements, and a deep network traffic discriminator is used for classification. The greedy algorithm is also used for optimized resource allocation. The proposed model is developed in Mininet with a Pox controller to prove Quality of Service (QoS) improvement paralleled with maximized utilization. The simulation results show that the proposed model has intelligent behaviour in network resource allocation based on flows' requirements compared to Spanning Tree Protocol (STP) and Dynamic and Adaptive Multipath Routing (DAMR) while maximizing the network utilization. [ABSTRACT FROM AUTHOR]

In industrial applications, mobile robot navigation involves autonomously moving towards specific target areas. Visual sensors are commonly used to identify and locate the target, enabling actions like manipulator grasping. However, when using the Yolo series for target positioning, two challenges arise. The first challenge is that the evaluation index Intersection over Union (IOU) only provides limited information and does not fully express the confidence in the target position. To overcome this, the Yolo layer utilizes a complete IOU as its loss function, which considers factors such as overlap, distance, and aspect ratio. Additionally, a Gaussian model layer is introduced between the non-maximum suppression layer and the output layer to estimate the coordinate uncertainty. The second challenge is that the prediction framework for target positioning lacks accuracy, resulting in significant deviations from the ground truth. To improve precision, this paper proposes combining the B-spline level set method for target segmentation. This technique adjusts the deviation from the minimum rectangular box surrounding the target. The proposed global optimization segmentation model, based on B-spline, effectively handles local optimal values in non-convex and traditional variational segmentation models. Experimental results demonstrate that: In the VOC dataset, compared with the Yolo-v3 method and Faster R-CNN method, the mAP of the GC Yolo-v3 method is increased by 2.95% and 5.85%, respectively, and the recall value is increased by 0.03 and 0.02, respectively. Compared with the Yolo-v3 method in VOC dataset, the average IOU value of the GC Yolo-v3 method is increased by 8.31%. In the real dataset, compared with the Yolo-v3 method and the Faster R-CNN method, the mAP of the GC Yolo-v3 method is increased by 3.32% and 7.41% respectively, and the recall value is increased by 0.08 and 0.05 respectively. The average IOU value of the GC Yolo-v3 method is 6.35% higher than that of the Yolo-v3 method in real dataset. [ABSTRACT FROM AUTHOR]

User reviews can play a big part in deciding a company's income in the e-commerce industry. Before making selections regarding any product or service, online users rely on reviews. As a result, the trustworthiness of online evaluations is vital for organisations and can directly impact their reputation and revenue. Because of this, some firms pay spammers to publish false reviews. Most recent studies to detect fake reviews utilise supervised learning. However, neural network techniques, a recent form of advanced technology, have been utilised extensively to detect fake reviews and have demonstrated their ability to do so. Thus, this paper first provides a benchmark study to analyse the performance of various machine learning algorithms with different feature extraction methods on five fake review datasets to present our results. Second, we propose three advanced language models for embedding reviews into the classifiers. Third, we conduct an exhaustive feature set evaluation study to find the best features in detecting fake reviews. Fourth, we analyse the performance of traditional machine learning, deep learning, and advanced deep learning models using different feature extraction methods on five fake review datasets. Finally, we integrate the ELECTRA model with CNN which can identify real or fake reviews. Our proposed technique utilises accuracy, precision, recall, and F1 score as assessment criteria to determine the leniency of the proposed model. For deep contextualised representation and neural classification, we integrate Single-Layer Perceptron (SLP), Multi-Layer Perceptron (MLP), and Convolutional Neural Networks (CNN) following the embedding layer of unique pre-trained models like ELMo, ELECTRA, and GPT2. The experimental results indicate that our proposed model outperforms state-of-the-art methods with improvements ranging from 1 to 7% in terms of the accuracy, F1 score. To the best of our knowledge, no prior work has evaluated such advanced pre-trained models' efficiency in detecting fake reviews. Further, this research comprehensively evaluates several machine-learning approaches and feature extraction strategies for fake online review detection. [ABSTRACT FROM AUTHOR]

Deep learning (LeCun et al. in Nature 521(7553):436–444, 2015) workloads are common in today's production clusters due to the proliferation of deep learning driven AI (Russell and Norvig, Artificial intelligence: a modern approach, Pearson Education, Upper Saddle River, 2003) services. As deep learning workloads and GPUs all grow more heterogeneous, efficient and fair resource scheduling are the key to the maximal performance of a deep learning cluster. However, existing cluster schedulers are largely not tailored to deep learning jobs, and typically specifying a fixed amount of resources for each job, prohibiting high resource efficiency and job performance. Thus we reconsider the problem in this paper, and to resolve them we proposes MixTran, which works fairly and efficiently for mixed deep learning workloads on heterogeneous GPU clusters. At first, MixTran abstracts heterogeneous GPU resources and distributes them fairly to users in the form of virtual tickets. Then, MixTran transfers uniform requests into a global optimization model, which can make efficient use of GPUs and satisfy the quantified resource requests, heterogeneous node constraints and user fairness constraints. Finally, MixTran conducts greedy resource trading according to the second trader price, which can benefit both users. In the evaluation, we prove MixTran can greatly decrease total execution time of deep learning workloads than traditional scheduler up to 30%–50% and maintain the fairness of multiple users. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoTs) is a complex and large network of all kinds of equipment and things that covers various areas such as military networks, industrial networks, smart urban infrastructures, education, and many other areas, high volume of traffic, and the growing number of connected equipment. to these networks has made intrusion detection in these networks a challenging issue; Therefore, protecting these networks from various attacks is of particular importance and can be established through planning and establishing effective security controls. One of these important and effective methods of security control is the use of an intrusion detection system (IDS). Due to the ability of machine learning and deep learning methods to learn from existing real examples, they are used to prevent, predict, and intelligently and quickly track complex and unknown future examples. The most important ability and feature of these methods are to detect and predict new attacks, which in many cases are mutated or inspired by previous attacks. Therefore, IoTs control and security systems use and develop new methods in creating safe and intelligent communication between their devices and equipment, which are activated using advanced deep learning and machine methods. Also, the success and advancements of deep learning methods in various fields and fields with extensive data have made activists in the field of the internet interested in these methods. In this paper, to solve the mentioned problems and challenges, we also present a new and effective method of detection based on long short-term memory (LSTM)–recurrent neural network (RNN) and kernel principal component analysis (PCA). To achieve data preprocessing, high accuracy detection rate, feature extraction, and attack detection is embedded in an end-to-end or global detection method. We have used the NSL-KDD dataset to evaluate the results and check the presented idea. The results of our experiments and simulations show that the proposed IDS-LSTM attack detection strategy and idea has better performance than several detection and prediction strategies that use neural networks, support vector machines (SVM). [ABSTRACT FROM AUTHOR]

Android is the most targeted mobile operating system for malware attacks. Most modern anti-malware solutions largely incorporate deep learning or machine learning techniques to detect malwares. In this paper, we conduct a comprehensive analysis on 10 deep learning and 5 machine learning classifiers in their abilities to identify Android malware applications. We used 1-gram dataset, 2-gram dataset and image dataset generated from the system call co-occurrence matrix for our experiments. Among the machine learning classifiers, XGBoost with 2-gram dataset showed the highest F1-score of 0.98. Also, the deep learning classifiers such as extreme learning machine with the system call images demonstrated the best F1-score of 0.952. We experimented using Gabor filters to investigate classifier performance on textures extracted from system call images. We observed an F1-score of 0.953 using the extreme learning machine with the Gabor images. We generated the Gabor image dataset by combining the images generated by passing system call images through 25 different Gabor configurations. In addition, to enhance the performance of the baseline classifiers, we considered the combination of autoencoders with machine learning classifiers. We observed that the amalgam of autoencoder with Random Forest displayed the best F1-score of 0.98. To evaluate the effectiveness of the aforesaid classifiers with diverse features on adversarial examples, we simulated a black-box based attack using a Generative Adversarial Network. The True Positive Rate of XGBoost on the 1-gram dataset, Random Forest on the 2-gram dataset and the Extreme Learning Machine on the system call image dataset significantly dropped to 0 from 0.98, 0.001 from 0.99 and 0 from 0.984 after the attack. Our experiments exposed a crucial vulnerability in classifiers used in modern anti-malware systems. A similar event in a real-world system could potentially render grave catastrophes. To defend against such probable attacks, we should continue further research and develop adequate security mechanisms. [ABSTRACT FROM AUTHOR]

The advanced development of communication technologies has made covert communications quite challenging to be recognized. By altering an entity's timing behavior during overt network communication, a covert timing channel (CTC) provides a way to leak sensitive data. This is becoming a serious threat due to the ability to transmit hidden messages without being detected by traditional security systems such as proxies and firewalls. In this paper, we explore the automatic classification and identification of covert timing channels using deep neural networks, namely, Long Short-Term Memory (LSTM), 1D-Convolutional Neural networks (1D-CNN), and a hybrid of LSTM and 1D-CNN. These classifiers have been trained and tested using sequence real inter-arrival times datasets labeled with covert or legitimate. The stream length differs for each dataset; (32, 64, 128, 256, and 512 bytes). Traditional machine-learning models have also been developed for comparisons and evaluation. The evaluation results showed that the hybrid of LSTM and CNN model outperforms other models either developed by deep neural networks or traditional machine learning with an accuracy of %97.5. [ABSTRACT FROM AUTHOR]

In recent years, serverless computing has emerged as a new cloud service model that provides an elegant computing framework for application development. It allows developers to focus solely on application logic without having to worry about the underlying infrastructure or execution environment. One of the most attractive features of serverless computing is scale-to-zero, where function instances are immediately released as soon as they complete their execution. This makes the computing model more cost-effective. However, when a new function request arrives, a container must be prepared from scratch, including initialization, configuration, and loading of required libraries and packages. This requires additional startup time, also known as cold-start delay, which can degrade the performance of the computing framework. This paper proposes an integrated model, called adaptive container provisioning model (ACPM), to reduce cold-start latency through runtime provisioning of containers. ACPM promises to reduce the frequency of cold-starts and delay due to cold-start, improving the overall performance of serverless execution. It consists of two phases. In the first phase, an efficient deep learning model, such as long short term memory (LSTM), is used to predict the required number of pre-warmed containers in advance, which can be served as soon as the function request arrives. This phase is integrated with a container placement module for faster delivery of a container to minimize the frequency of cold-start delay. The second phase is based on a sandboxing mechanism where groups of similar functions can be placed in a single container instead of creating a separate container for each incoming request. ACPM drastically reduces cold-start delay and improves overall performance. The effectiveness of ACPM has been extensively compared with standard approaches such as container-sharing, container-reuse, pre-warming, restore-based, etc. From the experiment, it can be observed that ACPM outperforms other existing models in reducing cold-start latency. [ABSTRACT FROM AUTHOR]

This paper introduces a novel and complete framework for solving different Internet of Things (IoT) applications, which explores eXplainable AI (XAI), deep learning, and evolutionary computation. The IoT data coming from different sensors is first converted into an image database using the Gamian angular field. The images are trained using VGG16, where XAI technology and hyper-parameter optimization are introduced. Thus, analyzing the impact of the different input values in the output and understanding the different weights of a deep learning model used in the learning process helps us to increase interpretation of the overall process of IoT systems. Extensive testing was conducted to demonstrate the performance of our developed model on two separate IoT datasets. Results show the efficiency of the proposed approach compared to the baseline approaches in terms of both runtime and accuracy. [ABSTRACT FROM AUTHOR]

Pyrolysis is one of the thermochemical methods used to produce value-added products from biomass. Thermogravimetric analysis (TGA) is frequently used to examine the energy potential and thermal behavior of biomass, coal, and their blends. The investigation of the TGA data using Artificial Neural Networks (ANN) is one of the most important research areas in recent years. While there are different research papers on the use of Machine Learning (ML) in this field, there is a lack of systematic application of deep learning (DL) algorithms. As such, we applied DL algorithms together with ML algorithms to evaluate the predictive performance of thermal behaviors of proposed bioenergy sources. Thermal behavior of tomato, pepper, eggplant, squash, and cucumber harvest wastes, the equal mass (20%) mixture of them, and the blends of the mixture with coal in the ratios of 20, 33, and 50% under nitrogen atmosphere were investigated by the TGA and ML models. Based on the pyrolysis thermal behavior of the harvest wastes, the eggplant, pepper, tomato, and 5-biomass mixture had the highest conversion potential. According to the thermal behavior of co-pyrolysis of coal and harvest waste mixtures, it had positive effects on pyrolysis conversion degrees and temperature range compared to the coal, and therefore, they can be used as alternative sources for energy production. The MSE and R 2 scores of Bi-directional LSTM demonstrate that an improved performance can be obtained with DL based solutions. Promising results were obtained when the Bi-directional LSTM is applied for modeling the pyrolysis. [ABSTRACT FROM AUTHOR]

Data analytics projects span all types of domains and applications. Researchers publish results using certain datasets and classification models. They present results with a summary of the performance metrics of their evaluated classifiers. However, readers and evaluators may not be able to compare results from the different papers for several reasons. One reason is the variations in the classification models and specific settings used in those models; a second reason is the variations of computing resources and environments used to produce those results. A third reason is a variation in the input datasets, preprocessing steps, etc. There is a need for benchmarks or baselines, where researchers can reuse the same benchmark specifications and be able to produce the same reported results. This paper aims to aggregate models from different research papers and apply them using the same classification settings, the same computing resources, and the same datasets and preprocessing steps. The goals are to show a comparable comparison and also to build a benchmark for future assessment and improvements. Four popular IoT security datasets are used in this paper. Different classification models are evaluated and are applied in the same datasets using the same settings. The classifiers results are reported and compared based on several popular performance metrics. [ABSTRACT FROM AUTHOR]

Scholarly big data refer to the rapidly growing scholarly source of information, including a large number of authors, papers, and massive scale scholarly networks. Extracting the profile attributes for Web users is an important step in Web user analysis. For Web scholarly users, profile attributes extraction should integrate multi-source and heterogeneous information resources. However, the traditional extraction models have two main drawbacks: (1) The traditional models require manual feature selection based on specific domain knowledge; (2) The traditional models cannot adapt to the diversities of Scholarly Web pages and cannot discover the relationships between different target entities which are far apart in different domains. To address these issues, we propose a profile attributes extraction model, PAE-NN, based on a Bi-LSTM-CRF neural network. This model can automatically extract the characteristics and contextual representations of each extracting entity through a Recurrent Neural Network with end-to-end training. It takes advantage of the long-memory sequence characteristics of LSTM network to effectively discover the long-term dependencies on extracting entities. Our experimental results on published datasets from the SMPCUP2017 Open Academic Competition and Aminer demonstrate that the proposed PAE-NN model outperforms existing models in terms of extraction precision, recall, and F1-score with large-scale training data. [ABSTRACT FROM AUTHOR]

Due to its automatic feature learning ability and high performance, deep learning has gradually become the mainstream of artificial intelligence in recent years, playing a role in many fields. Especially in the medical field, the accuracy rate of deep learning even exceeds that of doctors. This paper introduces several deep learning algorithms: Artificial Neural Network (NN), FM-Deep Learning, Convolutional NN and Recurrent NN, and expounds their theory, development history and applications in disease prediction; we analyze the defects in the current disease prediction field and give some current solutions; our paper expounds the two major trends in the future disease prediction and medical field—integrating Digital Twins and promoting precision medicine. This study can better inspire relevant researchers, so that they can use this article to understand related disease prediction algorithms and then make better related research. [ABSTRACT FROM AUTHOR]

The fast spread of the COVID-19 over the world pressured scientists to find its cures. Especially, with the disastrous results, it engendered from human life losses to long-term impacts on infected people's health and the huge financial losses. In addition to the massive efforts made by researchers and medicals on finding safe, smart, fast, and efficient methods to accurately make an early diagnosis of the COVID-19. Some researchers focused on finding drugs to treat the disease and its symptoms, others worked on creating effective vaccines, while several concentrated on finding inhibitors for the key enzymes of the virus, to reduce its spreading and reproduction inside the human body. These enzymes' inhibitors are usually found in aliments, plants, fungi, or even in some drugs. Since these inhibitors slow and halt the replication of the virus in the human body, they can help fight it at an early stage saving the patient from death risk. Moreover, if the human body's immune system gets rid of the virus at the early stage it can be spared from the disastrous sequels it may leave inside the patient's body. Our research aims to find aliments and plants that are rich in these inhibitors. In this paper, we developed a deep learning application that is trained with various aliments, plants, and drugs to detect if a component contains SARS-CoV-2 key inhibitor(s) intending to help them find more sources containing these inhibitors. The application is trained to identify various sources rich in thirteen coronavirus-2 key inhibitors. The sources are currently just aliments, plants, and seeds and the identification is done by their names. [ABSTRACT FROM AUTHOR]

Smart cities are being enabled all around the world by Internet of Things (IoT) applications. A smart city idea necessitates the integration of information and communication technologies and devices throughout a network in order to provide improved services to consumers. Because of their increasing amount and mobility, they are increasingly appealing to attackers. Therefore, several solutions, including as encryptions, authentication, availability, and data integrity, have been combined to protect IoT. Intrusion detection systems (IDSs) are a powerful security tool that may be improved by incorporating machine learning (ML) and deep learning (DP) techniques. This paper presents a novel intrusion detection framework for IoT-based smart environments with Ensemble Learning called IDS-SIoEL. Typically, the framework proposed an optimal anomaly detection model that uses AdaBoost, and combining different feature selection techniques Boruta, mutual information and correlation furthermore. The proposed model was evaluated on IoT-23, BoT-IoT, and Edge-IIoT datasets using the GPU. When compared to existing IDS, our approach provides good rating performance features of ACC, recall, and precision, with around 99.9% on record detection and calculation time of 33.68 s for learning and 0.02156 s for detection. [ABSTRACT FROM AUTHOR]

The K-complex is one of the most important and noticeable features in the electroencephalography (EEG) signal, therefore its detection is critical for EEG signal analysis. It is used to diagnose neurophysiologic and cognitive disorders as well as sleep studies. Existing detection methods largely depend on tedious, time-consuming, and error-prone manual inspection of the EEG waveform. In this paper, a highly accurate K-complex detection system is developed. Based on multiple convolutional neural network (CNN) feature extraction backbones and EEG waveform images, you only look once v3 (YOLOv3) detector was designed, trained, and tested. Extensive performance evaluation was performed using five deep transfer learning feature extraction models; Darknet-53, MobileNets, ResNet-18, SqueezeNet, and Darknet-53-coco. The dataset was comprised of 10948 images of EEG waveforms, with the K-complex location automatically annotated with bounding boxes. The Darknet-53 model performed consistently high (i.e., 89.84–99.44% precision and 10.41–0.55% miss rate). Thus, it is possible to perform automatic K-complex detection in real-time with high accuracy that aid practitioners in speedy EEG inspection. [ABSTRACT FROM AUTHOR]

Popular deep learning frameworks like PyTorch utilize GPUs heavily for training, and suffer from out-of-memory (OOM) problems if memory is not managed properly. CUDA Unified Memory (UM) allows the oversubscription of tensor objects in the GPU, but suffers from heavy performance penalties. In this paper, we build upon our UM implementation and create and utilize a minimal overhead CUPTI dynamic profiler to trace unified memory page fault and memory transfer statistics in PyTorch applications. We also implement CUDA memory prefetch and advise API which can be called directly from the PyTorch application based on the dynamically profiled statistics to improve oversubscription performance in various PyTorch models including Resnet and BERT. [ABSTRACT FROM AUTHOR]

Deep learning has been postulated as a solution for numerous problems in different branches of science. Given the resource-intensive nature of these models, they often need to be executed on specialized hardware such graphical processing units (GPUs) in a distributed manner. In the academic field, researchers get access to this kind of resources through High Performance Computing (HPC) clusters. This kind of infrastructures make the training of these models difficult due to their multi-user nature and limited user permission. In addition, different HPC clusters may possess different peculiarities that can entangle the research cycle (e.g., libraries dependencies). In this paper we develop a workflow and methodology for the distributed training of deep learning models in HPC clusters which provides researchers with a series of novel advantages. It relies on udocker as containerization tool and on Horovod as library for the distribution of the models across multiple GPUs. udocker does not need any special permission, allowing researchers to run the entire workflow without relying on any administrator. Horovod ensures the efficient distribution of the training independently of the deep learning framework used. Additionally, due to containerization and specific features of the workflow, it provides researchers with a cluster-agnostic way of running their models. The experiments carried out show that the workflow offers good scalability in the distributed training of the models and that it easily adapts to different clusters. [ABSTRACT FROM AUTHOR]

We have designed and implemented a new data processing framework called "MeLoN" (Multi-tenant dEep Learning framework On yarN) which aims to effectively support distributed deep learning applications that can show another type of data-intensive workloads in the YARN-based Hadoop ecosystem. MeLoN is developed as one of Hadoop YARN applications so that it can transparently co-host existing deep learning applications with other data processing workflows. In this paper, we present comprehensive techniques that can effectively support multiple deep learning applications in a Hadoop YARN cluster by leveraging fine-grained GPU over-provisioning policy and a high-performance parallel file system for data staging which can improve the overall system throughput. Through our extensive experiments based on the representative deep learning workloads, we demonstrate that MeLoN can make an effective convergence of deep learning and the big data platform Hadoop by employing YARN-based resource allocation and execution mechanisms for running distributed deep learning applications. We believe that MeLoN can bring many additional interesting research issues including profiling of expected GPU memory usages of deep learning applications, supporting more complicated deep learning related jobs based on queuing systems which can ultimately contribute to a new data processing framework in the YARN-based Hadoop ecosystem. [ABSTRACT FROM AUTHOR]

Recently, improving the overall resource utilization through efficient scheduling of applications on graphic processing unit (GPU) clusters has been a concern. Traditional cluster-orchestration platforms providing GPUs exclusively for applications constrain high resource utilization. Co-execution of GPU applications is suggested to utilize limited resources. However, the co-execution of GPU applications without considering their diverse characteristics can lead to their unpredictable performances owing to interference resulting from contention and unbalanced usage of resources among applications. This paper proposes an interference-aware execution framework with Co-scheML for various GPU applications such as high performance computing (HPC), deep learning (DL) training, and DL inference. Various resource-usage characteristics of GPU applications are analyzed and profiled to identify various degrees of their application interference. As interference prediction is challenging owing to the complexity of GPU systems, an interference model is generated by applying defined GPU metrics to machine learning (ML) models. A Co-scheML scheduler deploys applications to minimize the interference using the predicted interference from the constructed model. Experimental results of our framework demonstrated that the resource utilization improved by 24%, the average job completion time (JCT) improved by 23%, and the makespan shortened by 22% on average, compared to baseline schedulers. [ABSTRACT FROM AUTHOR]

Convolutional Neural Network (CNN) is one of the most widely used deep learning models in pattern and image recognition. It can train a large number of datasets and get valuable results. The deep Residual Network (ResNet) is one of the most innovative CNN architecture to train thousands of layers or more and leads to high performance for complex problems. This deep model trains the neural networks with the idea of identity shortcut connection that skips layers. In this paper, we built a hyperparameter optimization approach for the ResNet models with different numbers of layers and show that our optimization leads to significant performance improvements. We developed a manual search approach by enhancing the traditional data augmentation proposed by the previous approaches. Using the CIFAR-100 dataset for classification, our approach significantly improves the classification error rate. It also reduces the computing time compared with the previous automatic approaches. Our approach leads to 23.01% error rate and 4 h and 15 min of computing time of ResNet-164 for the CIFAR-100 dataset using the NIVIDIA Tesla P100 GPU. [ABSTRACT FROM AUTHOR]

Cyber Physical System (CPS) plays an important role in industry 4.0 applications such as smart factories, smart energy, smart transportation, smart buildings, smart healthcare, etc. Similarly, Cyber Physical Sensor System (CPSS) has gained popularity in recent times and is composed of a computing platform linked to an actuator, sensor, and wireless access point. In real-time scenarios, CPSS continuously gathers data from physical objects and conducts real-time control events based on the process algorithm. Then, the gathered data is transferred to the control centre or cloud services via network layer for further processing. In this scenario, there exists a need to identify the way of utilizing the intellect correctly, by designing effective data sensing and fusion schemes for CPSS. With this background, the current paper presents a Deep Learning with Metaheuristics based Data Sensing and Encoding (DLMB-DSE) scheme for CPSS. The aim of the proposed DLMB-DSE technique is to present a prediction-based data sensing and fusion approach to reduce the quantity of data communication and maintain maximum coverage by ensuring security. DLMB-DSE technique involves the design of Optimal Deep Belief Network (DBN) with Adagrad optimizer to primarily predict the data of the succeeding period with minimum number of data items. It also helps in making the primary predicted value, estimate the actual value, with maximum accuracy. Besides, Multi-Key Homomorphic Encryption (MKHE) technique is also applied for useful data encoding and decoding processes, thereby accomplishing security. Moreover, the novelty of the study lies in optimal key generation process, followed in MKHE technique, using Equilibrium Optimizer (EO). This helps in improving the security. A wide range of experiments was implemented to validate the better performance of the proposed DLMB-DSE technique. The experimental results exhibit the promising performance of DLMB-DSE approach over other methods under different measures. [ABSTRACT FROM AUTHOR]

Federated Learning (FL), Artificial Intelligence (AI), and Explainable Artificial Intelligence (XAI) are the most trending and exciting technology in the intelligent healthcare field. Traditionally, the healthcare system works based on centralized agents sharing their raw data. Therefore, huge vulnerabilities and challenges are still existing in this system. However, integrating with AI, the system would be multiple agent collaborators who are capable of communicating with their desired host efficiently. Again, FL is another interesting feature, which works decentralized manner; it maintains the communication based on a model in the preferred system without transferring the raw data. The combination of FL, AI, and XAI techniques can be capable of minimizing several limitations and challenges in the healthcare system. This paper presents a complete analysis of FL using AI for smart healthcare applications. Initially, we discuss contemporary concepts of emerging technologies such as FL, AI, XAI, and the healthcare system. We integrate and classify the FL-AI with healthcare technologies in different domains. Further, we address the existing problems, including security, privacy, stability, and reliability in the healthcare field. In addition, we guide the readers to solving strategies of healthcare using FL and AI. Finally, we address extensive research areas as well as future potential prospects regarding FL-based AI research in the healthcare management system. [ABSTRACT FROM AUTHOR]

Presently, Supervisory Control and Data Acquisition (SCADA) systems are broadly adopted in remote monitoring large-scale production systems and modern power grids. However, SCADA systems are continuously exposed to various heterogeneous cyberattacks, making the detection task using the conventional intrusion detection systems (IDSs) very challenging. Furthermore, conventional security solutions, such as firewalls, and antivirus software, are not appropriate for fully protecting SCADA systems because they have distinct specifications. Thus, accurately detecting cyber-attacks in critical SCADA systems is undoubtedly indispensable to enhance their resilience, ensure safe operations, and avoid costly maintenance. The overarching goal of this paper is to detect malicious intrusions that already detoured traditional IDS and firewalls. In this paper, a stacked deep learning method is introduced to identify malicious attacks targeting SCADA systems. Specifically, we investigate the feasibility of a deep learning approach for intrusion detection in SCADA systems. Real data sets from two laboratory-scale SCADA systems, a two-line three-bus power transmission system and a gas pipeline are used to evaluate the proposed method's performance. The results of this investigation show the satisfying detection performance of the proposed stacked deep learning approach. This study also showed that the proposed approach outperformed the standalone deep learning models and the state-of-the-art algorithms, including Nearest neighbor, Random forests, Naive Bayes, Adaboost, Support Vector Machine, and oneR. Besides detecting the malicious attacks, we also investigate the feature importance of the cyber-attacks detection process using the Random Forest procedure, which helps design more parsimonious models. [ABSTRACT FROM AUTHOR]

This paper introduces a novel framework for identifying a group of intrusions in the context of the Internet of Things (IoT). It combines both deep learning and decomposition. A set of data is first collected and a recurrent neural network is used to estimate the different individual intrusions. These individual intrusions are then used to identify a group of outliers based on a decomposition strategy. As case studies, the proposed solutions have been experimentally evaluated using two computer network intrusion datasets, namely (1) IDS 2018, and (2) LUFlow. The results show the benefits of the proposed framework and clear superiority in comparison to the state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

In recent years, the exponential growth of malware has posed a significant security threat to intelligent systems. Earlier static and dynamic analysis methods fail to achieve effective recognition rate and incurs high computational complexity. The recently developed machine learning (ML) and deep learning (DL) models can be employed to detect and classify cyberattacks and Malware efficiently. This paper presents a fusion of deep learning based cyberattack detection and classification model for intelligent systems named FDL-CADIS technique. The proposed FDL-CADIS technique transforms the Malware binary files into two-dimensional images, which are then classified by the fusion model. The FDL-CADIS technique employs the binary input images into the MobileNetv2 model for the extraction of features and the hyper parameter tuning process takes place utilizing the black widow optimization technique. The MobileNetv2 model derives all features from the Malware dataset and trains the model using the derived features. Finally, an ensemble of voting based classifiers, including gated recurrent unit and long short-term memory techniques, for Malware cyberattack detection and classification was developed. A comprehensive range of experimental analysis is performed against the benchmark dataset to demonstrate the FDL-CADIS technique's promising performance. According to the comparative analysis of the results, the FDL-CADIS technique outperformed current approaches. [ABSTRACT FROM AUTHOR]

This paper proposes a multichannel deep learning approach for lung disease detection using chest X-rays. The multichannel models used in this work are EfficientNetB0, EfficientNetB1, and EfficientNetB2 pretrained models. The features from EfficientNet models are fused together. Next, the fused features are passed into more than one non-linear fully connected layer. Finally, the features passed into a stacked ensemble learning classifier for lung disease detection. The stacked ensemble learning classifier contains random forest and SVM in the first stage and logistic regression in the second stage for lung disease detection. The performance of the proposed method is studied in detail for more than one lung disease such as pneumonia, Tuberculosis (TB), and COVID-19. The performances of the proposed method for lung disease detection using chest X-rays compared with similar methods with the aim to show that the method is robust and has the capability to achieve better performances. In all the experiments on lung disease, the proposed method showed better performance and outperformed similar lung disease existing methods. This indicates that the proposed method is robust and generalizable on unseen chest X-rays data samples. To ensure that the features learnt by the proposed method is optimal, t-SNE feature visualization was shown on all three lung disease models. Overall, the proposed method has shown 98% detection accuracy for pediatric pneumonia lung disease, 99% detection accuracy for TB lung disease, and 98% detection accuracy for COVID-19 lung disease. The proposed method can be used as a tool for point-of-care diagnosis by healthcare radiologists.Journal instruction requires a city for affiliations; however, this is missing in affiliation 3. Please verify if the provided city is correct and amend if necessary.correct [ABSTRACT FROM AUTHOR]

Federated Learning (FL) is a technology that facilitates a sophisticated way to train distributed data. As the FL does not expose sensitive data in the training process, it was considered privacy-safe deep learning. However, a few recent studies proved that it is possible to expose the hidden data by exploiting the shared models only. One common solution for the data exposure is differential privacy that adds noise to hinder such an attack, however, it inevitably involves a trade-off between privacy and utility. This paper demonstrates the effectiveness of image augmentation as an alternative defense strategy that has less impact of the trade-off. We conduct comprehensive experiments on the CIFAR-10 and CIFAR-100 datasets with 14 augmentations and 9 magnitudes. As a result, the best combination of augmentation and magnitude for each image class in the datasets was discovered. Also, our results show that a well-fitted augmentation strategy can outperform differential privacy. [ABSTRACT FROM AUTHOR]

Deep learning has gained huge traction in recent years because of its potential to make informed decisions. A large portion of today's deep learning systems are based on centralized servers and fall short in providing operational transparency, traceability, reliability, security, and trusted data provenance features. Also, training deep learning models by utilizing centralized data is vulnerable to the single point of failure problem. In this paper, we explore the importance of integrating blockchain technology with deep learning. We review the existing literature focused on the integration of blockchain with deep learning. We classify and categorize the literature by devising a thematic taxonomy based on seven parameters; namely, blockchain type, deep learning models, deep learning specific consensus protocols, application area, services, data types, and deployment goals. We provide insightful discussions on the state-of-the-art blockchain-based deep learning frameworks by highlighting their strengths and weaknesses. Furthermore, we compare the existing blockchain-based deep learning frameworks based on four parameters such as blockchain type, consensus protocol, deep learning method, and dataset. Finally, we present important research challenges which need to be addressed to develop highly trustworthy deep learning frameworks. [ABSTRACT FROM AUTHOR]

Road transportation is a statutory organ in a modern society; however it costs the global economy over a million lives and billions of dollars each year due to increase in road accidents. Researchers make use of machine learning to detect and predict road accidents by incorporating the social media which has an enormous corpus of geo-tagged data. Twitter, for example, has become an increasingly vital source of information in many aspects of smart societies. Twitter data mining for detection and prediction of road accidents is one such topic with several applications and immense promise, although there exist challenges related to huge data management. In recent years, various approaches to the issue have been offered, but the techniques and conclusions are still in their infancy. This paper proposes a deep learning accident prediction model that combines information extracted from tweet messages with extended features like sentiment analysis, emotions, weather, geo-coded locations, and time information. The results obtained show that the accuracy is increased by 8% for accident detection, making test accuracy reach 94%. In comparison with the existing state-of-the-art approaches, the proposed algorithm outperformed by achieving an increase in the accuracy by 2% and 3% respectively making the accuracy reach 97.5% and 90%. Our solution also resolved high-performance computing limitations induced by detector-based accident detection which involved huge data computation. The results achieved has further strengthened confidence that using advanced features aid in the better detection and prediction of traffic accidents. [ABSTRACT FROM AUTHOR]

Storage workload prediction is a critical step for fine-grained load balancing and job scheduling in realtime and adaptive cluster systems. However, how to perform workload time series prediction based on a deep learning method has not yet been thoroughly studied. In this paper, we propose a storage workload prediction method called CrystalLP based on deep learning. CrystalLP includes workload collecting, data preprocessing, time series prediction, and data post-processing phase. The time series prediction phase is based on a long short-term memory network (LSTM). Furthermore, to improve the efficiency of LSTM, we study the sensitivity of the hyperparameters in LSTM. Extensive experimental results show that CrystalLP can obtain performance improvement compared with three classic time series prediction algorithms. [ABSTRACT FROM AUTHOR]

Aiming at the three main problems of power system—leakage, high temperature and physical damage, a new image fusion of fault detection method in power system based on deep learning is proposed in this paper. The core of deep learning is achieved by capsule network model. The model is trained and tested by self-built image dataset of power system. There are three types of dataset: visible images,infrared images and ultraviolet images. After being preprocessed and feature-extracted, the visible image is used as the fusion image background, the infrared image provides the thermal information of power equipment, and the ultraviolet image provides the electric field information on the exterior of power equipment. The collected images are decomposed into corresponding high frequency component image and low frequency component image respectively, which reconstructed into fused images by the capsule network model. With the registration of the three types of images, the faults in the power system can be detected and displayed accurately in the fused image. [ABSTRACT FROM AUTHOR]

Software defined network (SDN) decouples the network control and data planes. Despite various advantages of SDNs, they are vulnerable to various security attacks such anomalies, intrusions, and Denial-of-Service (DoS) attacks and so on. On the other hand, any anomaly and intrusion in SDNs can affect many important domains such as banking system and national security. Therefore, the anomaly detection topic is a broad research domain, and to mitigate these security problems, a great deal of research has been conducted in the literature. In this paper, the state-of-the-art schemes applied in detecting and mitigating anomalies in SDNs are explained, categorized, and compared. This paper categorizes the SDN anomaly detection mechanisms into five categories: (1) flow counting scheme, (2) information-based scheme, (3) entropy-based scheme, (4) deep learning, and (5) hybrid scheme. The research gaps and major existing research issues regarding SDN anomaly detection are highlighted. We hope that the analyses, comparisons, and classifications might provide directions for further research. [ABSTRACT FROM AUTHOR]

Since deep learning is introduced into the field of super-resolution (SR), many deep learning-based SR methods have been proposed and achieved good results. At present, most neural networks use ordinary convolution and deeper neural layer in image super-resolution reconstruction in order to achieve better results. Therefore, most existing models have a massive amount of parameters and calculation, which limits the practical application. To solve this problem, in this paper, we propose an efficient feature cheap convolution SR (FCCSR) model, which consists of several multi-level information fusion Blocks (MIFB). In the MIFB, the features are extracted by combining normal convolution and cheap convolution, and the features obtained from each layer of convolution are fused. Under limited parameters and reconstruction speed, it enables the whole model to obtain the features from shallow to deep layers in the low-resolution (LR) images. Secondly, a novel lightweight channel attention module is designed to obtain fewer parameters and better performance. Finally, we add a gradient loss function to the original L1 loss function, making the model pay more attention to the high-frequency part of the LR image and making the reconstructed image's texture details more straightforward. The experimental results show that the proposed model has reached the most advanced level regarding image quality, memory consumption, parameter number, and calculation amount. The operation speed of FCCSR is roughly the same as that of the current STOA models, and the number of parameters is only 0.63 M. On some data sets, the best performance of FCCSR can be improved by 7%. [ABSTRACT FROM AUTHOR]

Recently, phishing attacks have become one of the most prominent social engineering attacks faced by public internet users, governments, and businesses. In response to this threat, this paper proposes to give a complete vision to what Machine learning is, what phishers are using to trick gullible users with different types of phishing attacks techniques and based on our survey that phishing emails is the most effective on the targeted sectors and users which we are going to compare as well. Therefore, more effective phishing detection technology is needed to curb the threat of phishing emails that are growing at an alarming rate in recent years, thus will discuss the techniques of mitigation of phishing by Machine learning algorithms and technical solutions that have been proposed to mitigate the problem of phishing and valuable awareness knowledge users should be aware to detect and prevent from being duped by phishing scams. In this work, we proposed a detection model using machine learning techniques by splitting the dataset to train the detection model and validating the results using the test data , to capture inherent characteristics of the email text, and other features to be classified as phishing or non-phishing using three different data sets, After making a comparison between them, we obtained that the most number of features used the most accurate and efficient results achieved. the best ML algorithm accuracy were 0.88, 1.00, and 0.97 consecutively for boosted decision tree on the applied data sets. [ABSTRACT FROM AUTHOR]

The huge cost of emergency situations could have fatal effects on humanity and society, and it could present a genuine threat to both of them. In fact, most people confronted with an emergency could feel psychological trauma, which will, for the most part, change over time as they can exhibit chaotic or even turbulent behaviours. The situation could worsen in the case of a pandemic as fear and anxiety invade and spread in addition to isolation and quarantine. In this paper, we propose to build a smart assistant, called SMAD, that could detect the symptoms of an emergency case as well as symptoms of a mental disorder while analysing the natural language speech of an ordinary citizen, during and after an emergency situation using natural language processing and deep learning sentiment analysis model to track the patient's mental state during an ongoing conversation. Our proposed smart assistant is an online human-bot interaction that could handle a variety of physical and mental circumstances of any emergency situation. The proposed approach is a smart healthcare service that consists of four interconnected modules: The information understanding module, the data collector module, the action generator module, and the mental analysis module, which is based on the sentiment analysis model performed on a social media dataset using a pre-trained word-embedding model. [ABSTRACT FROM AUTHOR]

Cloud computing (CC) offers various types of services for the users and it is also termed on-demand computing. Because of its increasing popularity, it is vulnerable to a variety of intruders who could compromise the integrity and privacy of data stored in the cloud. Due to its distributed nature, security is the most challenging one in the cloud solution. Privacy and security are the major problems in its victory of the on-demand service, but it is simply vulnerable to intruders for any kind of attack. To solve this problem, IDSs (intrusion detection systems) play a major task in identifying the threats on cloud infrastructure. This paper develops an efficient cloud IDS using the sandpiper-based feature selection and extended equilibrium deep transfer learning (EEDTL) classification to improve the overall security of a cloud-based computing environment. The number of features is reduced from the given intrusion dataset based on the sandpiper optimization algorithm (SOA) while maintaining the minimal loss of information. Finally, the EEDTL model is used for the classification of different attacks based on their selected optimal features. For fine-tuning the attributes in convolution layers, transfer learning uses a pre-trained network called AlexNet. Also, the extended equilibrium optimizer (EEO) is used to update the network weights. The proposed cloud IDS effectively classify whether the network traffic behavior is normal or attack. The proposed system is executed in python using the UNSW-NB15 dataset, and NSL-KDD dataset. The various evaluation metrics are used to show the efficiency of the proposed method and compared to the existing works. The simulation results show that the proposed method can able to detect intrusions with a high detection rate and a low false alarm rate (FAR) than other approaches. [ABSTRACT FROM AUTHOR]

Today, the high accuracy of deep learning has led to use in various domains such as image and voice classification. However, vast computations of deep neural networks (DNNs) have caused the inefficiency of traditional processors, resulting in the emergence of hardware accelerators. DNN accelerators have increased performance by exploiting opportunities such as data reuse and sparsity. In these accelerators, the dataflow is an essential factor, so that some of them have a reconfigurable architecture to support different mappings and dataflows. However, the accelerators explicitly designed for exploiting the data sparsity are usually non-reconfigurable and have fixed dataflow. This paper presents a new dataflow called Channel Dimension Stationary (CDS) for the MAERI (a Reconfigurable Neural Network Accelerator). It can be used for convolutional layers with sparse input feature maps (ifmaps). In the proposed dataflow, computations are based on the Cartesian product method. However, multiplications leading to useless results are avoided. To analyze the mapping based on CDS dataflow, we upgraded the mRNA tool (mapper for Reconfigurable Neural Accelerators), which includes an energy and performance analyzer of the mapping strategy in MAERI. By evaluation, we found that in the sparse ifmaps of 50%, 70%, and 90%, the proposed mapping on average can increase energy efficiency by 3x, 6x, and 13x respectively, without noticeable reduction of utilization. [ABSTRACT FROM AUTHOR]

Deep learning-based video anomaly detection methods have drawn significant attention in the past few years due to their superior performance. However, almost all the leading methods for video anomaly detection rely on large-scale training datasets with long training times. As a result, many real-world video analysis tasks are still not applicable for fast deployment. On the other hand, the leading methods cannot provide interpretability due to the uninterpretable feature representations hiding the decision-making process when anomaly detection models are considered as a black box. However, the interpretability for anomaly detection is crucial since the corresponding response to the anomalies in the video is determined by their severity and nature. To tackle these problems, this paper proposes an efficient deep learning framework for video anomaly detection and provides explanations. The proposed framework uses pre-trained deep models to extract high-level concept and context features for training denoising autoencoder (DAE), requiring little training time (i.e., within 10 s on UCSD Pedestrian datasets) while achieving comparable detection performance to the leading methods. Furthermore, this framework presents the first video anomaly detection use of combing autoencoder and SHapley Additive exPlanations (SHAP) for model interpretability. The framework can explain each anomaly detection result in surveillance videos. In the experiments, we evaluate the proposed framework's effectiveness and efficiency while also explaining anomalies behind the autoencoder's prediction. On the USCD Pedestrian datasets, the DAE achieved 85.9% AUC with a training time of 5 s on the USCD Ped1 and 92.4% AUC with a training time of 2.9 s on the UCSD Ped2. [ABSTRACT FROM AUTHOR]

This paper presents a comprehensive suite of techniques for optimized memory management in multi-GPU systems to accelerate deep learning application execution. We employ a hybrid utilization of GPU and CPU memories in a multi-GPU environment by effectively addressing contention issues in the shared interconnect (e.g., PCIe, NVLink). In addition, we designed and implemented an intelligent prefetching algorithm (from CPU memory to GPU) that achieves the highest processing throughput while sustaining a large mini-batch size. We successfully implemented our optimization techniques on TensorFlow, and performed extensive experiments in various multi-GPU environments including traditional PCIe and the latest high-bandwidth interconnect, NVLink. Evaluation results show that our proposed scheme actually improves computing performance by decreasing the I/O bottleneck, and effectively increasing the mini-batch size without sacrificing overall training throughput. [ABSTRACT FROM AUTHOR]