Social media has become integral to daily interactions and a key data source for researchers. Using COVID-19 as a case study, this work compares 24 social media datasets to address three research questions: 1) Is the dataset in compliance with the FAIR principles of being Findable, Accessible, Interoperable, and Reusable? 2) To what extent have people utilized social media to voice and exchange their apprehensions during the COVID-19 pandemic? 3) To what extent can social media datasets be utilized for natural language processing (NLP)-based COVID-19 pandemic surveillance? Leveraging the evaluation questions derived from the FAIR principles, the authors assess 24 social media datasets related to the COVID-19 pandemic. Additionally, they comprehensively analyze each dataset, including their composition, and the specific instances and features they encompass. They have initiated an attempt hoping that more researchers will join to create a data community where information can be repurposed and reused.

Software updates are essential to enhance security, fix bugs, and add better features to the existing software. While some users accept software updates, non-compliance remains a widespread issue. End users’ systems remain vulnerable to security threats when security updates are not installed or are installed with a delay. Despite research efforts, users’ noncompliance behavior with software updates is still prevalent. In this study, we explored how psychological factors influence users’ perception and behavior toward software updates. In addition, we investigated how information about potential vulnerabilities and risk scores influence their behavior. Next, we proposed a model that utilizes attributes from the National Vulnerability Database (NVD) to effectively assess the overall risk score associated with delaying software updates. Next, we conducted a user study with Windows OS users, showing that providing a risk score for not updating their systems and information about vulnerabilities significantly increased users’ willingness to update their systems. Additionally, we examined the influence of demographic factor, gender, on users’ decision-making regarding software updates. Our results show no statistically significant difference in male and female users’ responses in terms of concerns about securing their system. The implications of this study are relevant for software developers and manufacturers as they can use this information to design more effective software update notification messages. The communication of the potential risks and their corresponding risk scores may motivate users to take action and update their systems in a timely manner, which can ultimately improve the overall security of the system.

In the realm of precision medicine, the prediction of phenotypic traits from genotype data plays a pivotal role in understanding disease susceptibility, drug response, and overall health outcomes. Machine learning models trained on genotype data offer promising avenues for phenotype prediction. However, the black-box nature of these models often hinders interpretabil-ity, raising concerns about their reliability and trustworthiness in clinical decision-making. This paper explores the application of SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) on our existing CNN model to identify important features from genotype that contribute to phenotype prediction in yeast data. Both techniques on machine learning models include improving model transparency and interpretability by providing insights into prediction mechanisms. LIME focuses on local interpretability, explaining individual predictions tailored to specific instances swiftly, while SHAP computes feature importance using Shapley values, offering a comprehensive understanding of feature contributions. Our dataset contains 4,390 samples, each with 28,220 features in genotype which are used in the training, testing and validation of our existing CNN model. In our analysis of 20 phenotypes, we identify the number of relevant important features using SHAP, LIME methods, and the intersection which not only improves the Mean Squared Error (MSE) values compared to using all the features in the genotype data, but also the model training time.

As social media and online social networking sites have emerged as essential channels for everyday social interactions, researchers are increasingly turning to conversational social media as a valuable alternative source of insights into public sentiment and behavior during times of crisis. In a similar vein, our study investigates the utility of using conversational social media data for analyzing public sentiment related to the COVID-19 pandemic. Our research question is, to what extent can machine learning models based on conversational social media data help gauge public sentiment during a pandemic? To evaluate the performance of various machine learning algorithms in sentiment analysis tasks, we used a Twitter dataset and compared their effectiveness using eight different models (i.e., Decision Tree (DT), Random Forest (RF), Naïve Bayes (NB), Extreme Gradient Boosting (XGBoost), Logistic Regression, Support Vector Machine (SVM), Long Short-Term Memory (LSTM), and Bidirectional Encoder Representations from Transformers (BERT)). The results show that among the eight models, the top three models are BERT, LSTM, and SVM, achieving overall accuracies of 87.44%, 81.94%, and 80.02%, respectively. Employing sentiment analysis models on pandemic-related social media posts allows us to gain insights into of human emotions and sentiments during public health crises like a pandemic.

Approximately 65% of the organizations in the United States have fallen victim to a successful phishing attack. Many organizations offer anti-phishing training to their employees to defend against phishing attacks. The purpose of this study is to examine factors impacting the effectiveness of anti-phishing training and study the relationship between personality traits and phishing susceptibility. Participants filled out pre- and post-training surveys that included questions on identifying phishing and legitimate URLs and questions to determine DISC (Dominant, Influence, Steadiness, and Conscientiousness) personality traits. An analysis of the survey data shows that the participants’ average accuracy in detecting phishing URLs increased 8% (t = 2.144, p-value = 0.0374) and their confidence in their answer choices increased 6% (t = 2.032, p-value = 0.0464) from pre-training to post-training surveys. Before and after training, participants with the Influence personality trait had the lowest susceptibility while both Dominant and Steadiness personalities had the highest susceptibility before and after training respectively.

In Multi-label (ML) classification, each instance is associated with more than one class label. Incorporating the label correlations into the model is one of the increasingly studied areas in the last decade, mostly due to its potential in training more accurate predictive models and dealing with noisy/missing labels. Previously, multi-label learning classifier systems have been proposed that incorporate the high-order label correlations into the model through the label powerset (LP) technique. However, such a strategy cannot take advantage of the valuable statistical information in the label space to make more accurate inferences. Such information becomes even more crucial in ML image classification problems where the number of labels can be very large. In this paper, we propose a multi-label learning classifier system that leverages a structured representation for the labels through undirected graphs to utilize the label similarities when evolving rules. More specifically, we propose a novel scheme for covering classifier actions, as well as a method to calculate ML prediction arrays. The effectiveness of this method is demonstrated by experimenting on multiple benchmark datasets and comparing the results with multiple well-known ML classification algorithms.

Ransomware is a major malware attack experienced by large corporations and healthcare services. Ransomware employs the idea of cryptovirology, which uses cryptography to design malware. The goal of ransomware is to extort ransom by threatening the victim with the destruction of their data. Ransomware typically involves a 3-step process: analyzing the victim’s network traffic, identifying a vulnerability, and then exploiting it. Thus, the detection of ransomware has become an important undertaking that involves various sophisticated solutions for improving security. To further enhance ransomware detection capabilities, this paper focuses on an Application Programming Interface (API)-based ransomware detection approach in combination with machine learning (ML) techniques. The focus of this research is (i) understanding the life cycle of ransomware on the Windows platform, (ii) dynamic analysis of ransomware samples to extract various features of malicious code patterns, and (iii) developing and validating machine learning-based ransomware detection models on different ransomware and benign samples. Data were collected from publicly available repositories and subjected to sandbox analysis for sampling. The sampled datasets were applied to build machine learning models. The grid search hyperparameter optimization algorithm was employed to obtain the best fit model; the results were cross-validated with the testing datasets. This analysis yielded a high ransomware detection accuracy of 99.18% for Windows-based platforms and shows the potential for achieving high-accuracy ransomware detection capabilities when using a combination of API calls and an ML model. This approach can be further utilized with existing multilayer security solutions to protect critical data from ransomware attacks.

Internet-of-Things (IoT) technology has revolutionized our daily lives in many ways-whether it is the way we conduct our day-to-day activities inside our home, or the way we control our home environments remotely. Unbeknownst to the users, with the adoption of these “smart home” technologies, their personal space becomes vulnerable to security and privacy attacks. We conducted studies of vulnerabilities and security posture of smart home IoT devices. We started with a literature review on known vulnerability studies of the IoT devices, considering four categories of attacks: 1) physical; 2) network; 3) software; and 4) encryption. We then conducted our own vulnerability experiments that compared security postures between well known and lesser known vendors through misuse and abuse case analysis, followed by a review of coverage in major vulnerability databases. Based on our analysis, the main finding was the need for a stronger focus on the security posture of lesser known vendor devices as they are often less regulated and faceless scrutiny.

The popularity of Web-based computing has given increase to browser-based cyberattacks. These cyberattacks use websites that exploit various web browser vulnerabilities. To help regular users avoid exploit websites and engage in safe online activities, we propose a methodology of building a machine learning-powered predictive analytical model that will measure the risk of attacks and privacy breaches associated with visiting different websites and performing online activities using web browsers. The model will learn risk levels from historical data and metadata scraped from web browsers.

Several studies have examined ethical and privacy concerns in Human-Computer Interaction (HCI) and cybersecurity research. However, the approaches to assure proper ethics and privacy standards in cybersecurity research are not adequately studied. This paper introduces a framework to evaluate the ethical considerations of cybersecurity research studies. Our framework was used to evaluate the ethical and privacy considerations for the technical papers published in the proceedings of SOUPS 2017. Our research provides future researchers with methods for conducting ethically sound cybersecurity research.

In smart homes, users do not have enough options to express their privacy preferences and decide who can see their data, when their data can be used, and which part of data they want to share and which part they do not want. We propose a framework that uses machine learning algorithms to determine for what purpose, to whom and with what level of details the information will be shared.

Inspired by the human immune system, we explore the development of a new Multiple-Detector Set Artificial Immune System (mAIS) for the detection of mobile malware based on the information flows in Android apps. mAISs differ from conventional AISs in that multiple-detector sets are evolved concurrently via negative selection. Typically, the first detector set is composed of detectors that match information flows associated with malicious apps while the second detector set is composed of detectors that match the information flows associated with benign apps. The mAIS presented in this paper incorporates feature selection along with a negative selection technique known as the split detector method (SDM). This new mAIS has been compared with a variety of conventional AISs and mAISs using a dataset of information flows captured from malicious and benign Android applications. This approach achieved a 93.33% accuracy with a true positive rate of 86.67% and a false positive rate of 0.00%.

Following the popularity of Wikipedia, community authoring systems are increasingly in use as content sharing outlets. As such, a Web-based portal for sharing of user-generated content (e.g., course notes, quiz answers, etc.) shows prospect to be a great tool for social E-Learning. Among others, students are expected to be active contributors in such systems in order to offer and receive peer-help. However, privacy and territoriality concerns can be potential barriers to wide adoption of such technology. Understanding the preference for sharing learning content is the first step to address privacy and territoriality concerns of content providers. The authors conduct a survey among students in four university courses in order to learn their preference for sharing notes and quiz answers with three target groups: instructor, peer, and stranger (i.e., someone outside their class). The authors also examine the preference for acceptable method of sharing by inquiring about three methods: “anonymous sharing,” “pseudonymous sharing,” and “sharing with name”. They further investigate the importance of “content type,” “sharing method,” and “accessor type” on the preference for sharing. The survey also reveals respondents' self-reported reasons for controlling access to their generated learning content. The survey data indicate that even though the respondents have various levels of concerns, almost all of them are willing to share. The authors observe relationships between content type and respondents' preference over each of these parameters: accessor type, commentator type, and sharing method.

The goal of an intrusion detection system (IDS) is to monitor anomalous activities and differentiate between normal and abnormal behaviors (intrusion) in a host system or in a network. The IDS must maintain a high intrusion detection rate (DR) while simultaneously maintain a low false alarm rate (FAR). A high detection rate is the focus of this paper. In this paper, we implemented an Evolutionary General Regression Neural Network (E-GRNN) as a two-class classifier for intrusion detection based on features of application layer protocols (e.g., http, ftp, smtp, etc.) used in simulated network traffic activities. The E-GRNN is an evolutionary search-inspired General Regression Neural Network, which extracts the most salient features to reduce computational complexity and increase accuracy. Our research shows that the E-GRNN classifier was able to achieve a DR of 95.53% and an FAR of 2.11%.

Cloud technology can be used to support costeffective, scalable, and well-managed healthcare information systems. However, cloud computing, particularly multitenancy, introduces privacy and security issues related to personal health information (PHI). In this paper, we designed ontological models for healthcare workflow and multi-tenancy, and then applied HIPAA requirements on the models to generate HIPAA-compliant access control policies. We used Semantic Web Rule Language (SWRL) to represent access control policies as rules, and we verified the rules with an OWL-DL reasoner. Additionally, we implemented HIPAA security rules through access control policies in a cloud-based simulated healthcare environment. More specifically, we investigated access control policy specification and enforcement for cloud based healthcare information systems using an open source cloud platform, OpenStack. The results manifest HIPAA compliance through authorization policies that are capable of addressing vulnerabilities of multi-tenancy.

One of the ways a malicious router can launch a Denial of Service (DoS) attack is by rerouting IP-packets of other destinations to the victim node. In this paper, based on the observed traffic anomalies, we ropose using a Markov chain model to calculate trustworthiness of routers in order to isolate

IP-spoofing attacks remain one of the most damaging attacks in which an attacker replaces the original source IP address with a new one. Using the existing attacking tools to launch IP spoofing attacks, an attacker can now easily compromise access routers and not only the end-hosts. In this paper, we propose a trust-based approach using a Bayesian inference model that evaluates the trustworthiness of an access router with regards to forwarding packets without modifying their source IP address. The trust values for the access routers is computed by a judge router that samples all traffic being forwarded by the access routers. The simulation results show that our approach effectively detects malicious access routers. The results also show that our approach has a low impact on the network performance when no attack is present, and that it introduces little overhead traffic.

Trust is important in Ad-hoc networks because collaboration and cooperation among nodes are critical towards achieving the system's goals, such as routing reliability. In this paper, we seek to provide an understanding of the process of reputation-based trust approaches and related issues as they apply to routing in ad-hoc networks. We discuss the concept and properties of trust. We survey the following issues: the way wireless ad-hoc nodes first assume trust; the evidences that are collected to calculate trust; the calculations that are performed over the evidences; the way the decision is made to determine a trusted node; and how trust is updated to maintain a trusted environment. We provide a comparison of three well-established approaches and discuss some potential attacks against reputation-based trust.