Since the 21st century is the century of information technology, there are now a large number of Internet users, and the number of these users is steadily growing. Most people know the useful aspects of this network, but there are dangerous aspects of the Internet, which are reviewed in this paper. Today there are many types of threats such as phishing attacks, spam messages, malware, worms, spyware, Trojans, rooters, botnets. The main principle of protection against these threats is information and cyber literacy of using the Internet. The main purpose of this article is a summary of Internet threats based on a review of articles by previous researchers, as well as the definition of types of threats and attacks. As a result of the study, threats were filtered into two main types: threats of social engineering, where the main emphasis is on the information carrier, and technical threats, where various methods, algorithms and software implementations are used to hack directly into a computing device containing information.
Federated learning is proposed as a typical distributed AI technique to protect user privacy and data security, and it is based on decentralized datasets that train machine learning models by sharing model gradients rather than sharing user data. However, while this particular machine learning approach safeguards data from being shared, it also increases the likelihood that servers will be attacked. Joint learning models are sensitive to poisoning attacks and can effectively pose a threat to the global model when an attacker directly contaminates the global model by passing poisoned gradients. In this paper, we propose a joint learning poisoning attack method based on feature selection. Unlike traditional poisoning attacks, it only modifies important features of the data and ignores other features, which ensures the effectiveness of the attack while being highly stealthy and can bypass general defense methods. After experiments, we demonstrate the feasibility of the method.
In a 5G/B5G (Beyond 5G) network, Service level agreement (SLA), network efficiency, and service management are key issues for a network provider. A user equipment requests services (UE) are based on a key performance indicator (KPI) and key quality indicator (KQI) while selecting the network slice. Earlier predicting the benefits of 5G/B5G will be helpful for the service provider to improve the quality of service (QoS). Therefore, we aim to build a data-driven predictive application to screen the multiple services of 5G/B5G. In this context, multi-classification supervised machine learning models are applied to the publicly available dataset to classify the services of 5G/B5G. We performed different simulations with the ML algorithm, first with all features (KPI and KQI parameters), second with alone KPI, and finally with features selection methods. The multiclass Decision jungle (MDJ) model shows better performance in terms of accuracy of 90%, precision, and recall. Moreover, an application programming interface (API) of the MDJ model implemented and deployed in https://predictor5g.herokuapp.com/ and source code are available at https://github.com/tabrejmsc/5GServicePredictor.
The most popular method for analysing a text is sentiment analysis. It is quite helpful for social media monitoring since it enables us to get a broad sense of what the general population thinks about particular issues. However, it is also handy for business analysis and various other situations in which the text needs to be analyzed. In this paper we present a solution up to 90% correctly classify whether a tweet is associated with a disaster. Using a multi-detector ensembled approach, we create a reliable model that can predict accurately in situations powered by multiple CNNs and BERT detectors.
In recent decades, we have seen significant advancement in fields like Artificial Intelligence, Machine Learning, and Deep Learning, resulting in the developing of new technologies such as deepfake. Deepfakes are a form of digital media that replaces one identity’s likeness with another or creates a synthetic personality: in the form of high-quality realistic fake video, image, or audio. Deepfakes can be helpful in education, art, activism, and self-expression; however, some subjects can use deepfakes to harm the portrayal of people, create pornographic content, and spread misleading information. High-quality deepfakes are easy to build but incredibly difficult to detect, creating a need to explore technologies which can be helpful in deepfake detection. Therefore, we present a comparative study of deep-learning models that can benefit deepfake detection. We have explored four deep-learning models, namely, VGG16, MobileNetV2, XceptionNet, and InceptionV3 and trained these models on the FaceForensics++ dataset. Finally, we evaluate the performance of these models for deepfake detection and conclude the study with our observations and future scope for improvement in this field.
Consumption of energy accounts for a significant portion of the costs associated with carrying out any activity. Cloud computing has seen spectacular development in network technology because of the advancements in communication technology, the expansion of Internet usage, and the solution of issues on a bigger scale. The NP-hard challenge of scheduling and allocating resources to meet the needs of users in the Cloud has been referred to as a scheduling problem. For this reason, balancing the load is critical to overcoming all the issues listed above. Researchers have compared a wide range of load balancing techniques in the cloud to improve different performance metrics and help load balancing. The study also describes relation between green cloud computing and load balancing.
A primary concern in the development of mobile applications is ensuring that the apps can manage a rising user base without degrading the user experience or performance. Therefore, the application’s architecture needs to be able to handle several requests per minute. Considering the scalability aspect, we developed a multilingual Chatbot using RASA as the Natural Language Processing (NLP) library, flutter as the cross-platform application development framework, and Django framework for doing server-side manipulations. Our application showcases the best architectural and security practices in application development to make the model scalable over time and keep it free from any security threats. The entire work is divided into two main components: a mobile application for the chatbot and a desktop website where users can enter data in several languages to train a deep learning-based model for intent detection. The developed chatbot, designed to be a university bot, will automatically identify the user’s preferred language and provide responses in that language. We aimed to make the application user-friendly and as safe as possible.
Regulation of false news causes challenges in the dissemination of accurate information and leads to misconceptions that endanger national cohesion and individual tranquility. Because information fabricates how the folk encapsulates the world, it is prime to combat this unauthentic news. People establish their own ideas in addition to basing critical judgments on these news articles, thus inaccurate news can exert a disastrous impact on the culture of a society making segmentation of publication pieces as spurious or authentic highly important. Numerous academicians are endeavoring to spot spurious news, and Machine Learning has been shown to be fruitful. In this paper, varied Machine Learning Algorithms are utilized to generate models to classify a particular news piece as authentic or fake. Python was applied as the scripting language throughout this development. Individual Machine Learning Algorithms such as K Nearest Neighbors and Decision Trees along with in-built ensembled classifiers (Random Forest, Gradient Boosting) and custom ensembled models (Stacking, Maximum Voting Classifier) are pattered for the purpose. With the avail of the appropriate models and tools, the challenging work of detecting false news may be made simple. This paper has been able to achieve an accuracy of 91.5% in classifying news as true or fake by stacking three individual Machine Learning Models namely, K Nearest Neighbors, Support Vector Classifier and Logistic Regression into a custom-ensembled model.
Reading massive amounts of user-generated text and pulling out the relevant aspects and opinions is a complicated process. Summaries, on the other hand, help busy people who only have a little bit of time to read get the gist of the information quickly. Text summarization is the process of taking the original text and making a shorter version of it that still has all of its informational value and main idea. Humans have a hard time summarizing long texts by hand. Different ways to summarize a text can be put into groups based on the more general techniques of extractive and abstractive summarization. The research paper discusses the need for generating aspect-based summaries and Sentiment analysis. A framework is proposed based on extracting coherent aspects from the reviews and applying the extractive summarization method to generate summaries. In addition, providing insights into the reviews of tourist attractions by using aspect-based sentiment analysis. The results are evaluated using crowdsourcing, Fairsumm, and Centroid method. The crowdsourcing method gives the best result on aspect-based summaries.
To ensure that the software satisfies all necessary criteria, software testing is done. Regression Testing is an important part of software testing as it plays a crucial role in software testing procedure. It's done to make sure that everything works as it should, such that test cases with greater potential to discover faults are given higher priority. Test case selection is introduced to choose the most effective subset of test cases from a test suite, test case selection reduces the total cost, time, and effort necessary in the process of software testing. It works by removing the obsolete and redundant test cases. Within all the computational intelligence algorithms the nature inspired algorithms are highly demanded due to their increased efficiency in solving complex problems. Nature Inspired Approaches refers to the approaches derived directly from the nature. Nature acts as the main branch of knowledge that provides various solution for the complex problems. Incorporation of Nature Inspired Approaches in test case selection problem generates a subset of test cases from the test suite resulting in time, cost and effort reduction. With the use of research questions, this paper offers a thorough review of Test Case Selection methods utilizing Nature Inspired Approaches.
Blockchain is the most cutting-edge technology nowadays. Most organizations are trying to achieve decentralization using blockchain technology. But storing information in such a system incurs some charges due to the creation of new blocks which results in gas fees, and requires some processing which takes some time. Each transaction that executes in the blockchain uses some gas that the user must pay. Frequent updation/creation of data results in the creation of a new block resulting in gas fees, to reduce this transaction cost and the time to process the data, the user can cache multiple data in memory and push all the data in the blockchain in a single transaction which results in less gas usage as compared to pushing each data individually. The author has created a cachingmechanism to store multiple details in the memory and the user can push these stored data into the blockchain in a single click. In this implementation, the author has used a local blockchain provided by ganache which works on a local host. A dialogue box is shown to the user after a successful transaction showing the transaction and time taken to push the data to the blockchain. The gas fees and the current block number can be viewed in the ganache console.
Kidney Disease is the 3 rd most substantial cause as mentioned and has the highest rate of morbidity and mortality rated all over the world. Chronic Kidney Disease is detected at a very later stage therefore making it one of the most harmful diseases. According to the Electronic Medical Records it is very hard for the detection of Chronic Kidney Disease. Severe medical treatment like hemodialysis is required for the procedures conducted for the curing of Renal Disease. Chronic Kidney disease has affected nearly about 10% of the present day population which itself claims how contagious this disease is. Blood Pressure, Heart Ailment, Glucose Level, Sodium, Potassium are all the causes for the identification of CKD.
Because India’s economy has shrunk to a low level during COVID-19, building an emergency decision support model (EDSM) for economic growth factors is the main objective of this study. We develop the TODIM-VIKOR method under Pythagorean fuzzy information. For dealing with comparison problems, the Pythagorean fuzzy scoring function is presented. We also include a new entropy metric for assessing the degree of fuzziness in PyFS. We also present a new Jensen Shannon divergence metric for PyFS that can be used to compare the discrimination information of two PyFSs. In this article, we introduced entropy and divergence measures to derive objective weight in the TODIM-VIKOR approach. Establishes a novel emergency decision making (EDM) strategy under the Pythagorean fuzzy atmosphere, using economic growth considerations. We used TODIM to determine the overall dominance degree, which takes into account the bounded rationality of decision makers, and VIKOR to calculate the compromise ranking of alternatives.
This paper proposes a reverse auction and blockchain-based electric vehicles (EVs) energy trading (ET) approach by adopting the smart grid. EVs can be classified into prosumer or consumer during the ET in the proposed approach. We further introduced the Interplanetary File System (IPFS)-based cost-efficient ET approach considering the sixth-generation (6G) wireless network to address the scalability and response time issues of the data transactions. Furthermore, the proposed approach leverages a reverse auction mechanism for optimal ET for EVs. The reverse auction mechanism maximizes the profit for EVs participating in the ET scheme using smart grid. Based on the reverse auction mechanism, a consumer can choose to trade energy with the prosumer or smart grid based on solar panel efficiency. Finally, the proposed approach is simulated considering the various performance metrics such as transaction efficiency, profit for consumers, and convergence to enable an optimal and efficient ET system.
The airline industry has become a very organized sector given the magnitude of travel undertaken by the population these days. However, the airline ticket pricing is a fluid parameter because of its dependency on various factors. The variation in ticket prices has led to customers traveling with different fares even on the same flight. This results in revenue losses for the airline industry due to operation of under booked flights and high customer dissatisfaction. The prediction of flight prices at an early stage would help airlines strategize their operations and gather the necessary resources impacting a specific market segment level for a route. Thus, this paper aims to propose a model which can capture the variability between different factors affecting the cost of a flight ticket and arrive at the model with least mean absolute error for estimating the prices of a trip. The dataset used in this paper has been provided with prices of flight tickets for various airlines between March to June 2019. From the algorithms implemented on the dataset, an ensemble of the Artificial Neural Networks, XGBoost Regressor, and Light Gradient Boosting Machine Regression models gives the least mean absolute error of 1226.6989.
Sentiment analysis in social media content gained popularity over the past several years due to its various uses in understanding human computer interaction, understanding consumers behavior, psychology, smart system, etc. Due to the massive volumes of data available via social media, which is commonly used for expressing opinions and ideas, this issue has drawn a lot of attention. To identify emotions, this research will make use of a labelled Hinglish dataset. Deep learning-based techniques are utilized to identify emotions in tweets with mixed Hindi-English coding by utilizing transformer-based models and multilingual word embeddings derived from FastText methods. Various deep learning models, such as convolutional neural networks (CNN), long short term memory (LSTM), and bi-directional long short term memory (Bi- LSTM) have been used to analyze sentiments. In comparison to other models, the convolutional neural network (CNN) achieved the highest accuracy, 75.25%.
With the growth of the internet came various websites fulfilling a variety of purposes. This new medium for information sharing became broader with each day and in 1997, the world saw the birth of internet’s first video hosting site, "ShareYourWorld.com". Over the course of time, video sharing platforms have evolved into more complex systems that involve more than just video uploads and downloads. Today users can publish, share their own videos to the Internet via online video services which may be intended for specific users on a shared network or for viewing by the broader public. Recent developments have revealed that the comment section of these video sharing platforms have frequently being used to promote spam and vulgar language. In this paper, we have compared different techniques to find out the best one for each identified module of the video sharing platform.
In recent years, people got more interested in the research of the evolution of human-computer interaction (HCI). Right from the transition of a text-based command line interface to a graphical user interface (GUI) that uses a mouse, many changes have been observed. The ability to manipulate the mouse with hand movements is one of the areas of interest here. At present, due to the cluttered backgrounds and poor lighting, this is a complicated issue. The application demands extremely specific detection and recognition. In this article, a computer vision-based method has been shown for controlling a computer mouse with hand gestures in front of a linked camera. The user may move the mouse and use the left as well as right click functions using this systematic program.
Deafness and mutism tend to place those individuals in the limelight owing to their inability to hear or communicate with the outside world, attempting to distinguish them from the rest of society. The communication is mostly done in ASL (American Sign Language) or ISL (Indian Sign Language), with various motions signifying a certain letter, number, or phrase. Not all deaf people utilize sign language; some use hearing aids; nevertheless, few can afford a hearing device. All these issues are addressed by sign language. It is more efficient, simple, and adaptable; hence, the research focuses on creating a system for hand gesture recognition that delivers the message of hope for the entire society and its significance. The goal of this paper is to detect the sign gesture using deep learning and to improve existing systems to attain a efficient software with proficient and prominent software.
