While Federated learning (FL) is considered privacy-preserving by nature, it remains vulnerable to many attacks, such as data and model poisoning, that compromise data integrity and model accuracy. Conventional privacy-preserving federated learning (PPFL) mechanisms, including homomorphic encryption (HE), secure aggregation, and secure multiparty computation (SMPC) demonstrate several limitations, such as high computational complexity, significant communication overhead, and scalability challenges. To overcome the aforementioned issues, we propose an end-to-end secure FL architecture that integrates differential privacy (DP), zero-knowledge proof (ZKP), and median aggregation. DP prevents data leakage during model updates by introducing Laplacian noise for privacy preservation. ZKP is implemented through Schnorr’s protocol, which enables lightweight and efficient client authentication without revealing sensitive information. Finally, median aggregation is incorporated to mitigate the impact of outliers and adversarial updates, ensuring robust prediction aggregation. The experimental results indicate that the proposed approach outperforms other well-known PPFL methods including partially homomorphic encryption (PHE), fully homomorphic encryption (FHE) and SMPC. It delivers substantial improvements in global accuracy, especially for larger client counts, with gains of 10%-30% over the other methods. The client training time is significantly reduced by 70%-90%, ensuring faster processing. The approach also excels at reducing average round latency by 80%-95%, enhancing the overall efficiency of the system. Communication overhead is significantly reduced by 65%-85%, lowering data transfer costs per round. Furthermore, the size of the model is minimized by 60%-85%, making it more resource efficient and scalable for larger deployments.

The Internet of Things (IoT) has garnered significant attention from both industries and the research community. The diverse nature of IoT devices makes them a prime target for cybercriminals. An intelligent intrusion detection system (IDS) can quickly identify multiple types of cyberattacks within an IoT system. However, operational efficiency and safety become challenging issues when managing limited labeled data in IoT networks. This article proposes a novel cyberattack detection scheme using few-shot learning (FSL). The proposed scheme employs a self-attention mechanism with a Siamese network that learns to recognize normal and malicious network traffic patterns through FSL. The Siamese network architecture facilitates efficient sample comparison by learning a shared representation. Incorporating an attention mechanism further enhances its ability to focus on discriminative features, improving attack detection accuracy for rapidly evolving intrusion patterns. The effectiveness of the proposed framework is evaluated through extensive experiments on the latest Edge-IIoTset dataset. The experimental findings demonstrate that the proposed IDS achieved the best accuracy of 99.75% with the complete dataset. In a few-shot performance evaluation, the designed architecture achieved the best accuracy of 78.69%, 81.19%, and 81.83% for 1 shot, 5 shots, and 10 shots, respectively.

This paper proposes an integrated approach called FL-RSD, leveraging the key advantages of Federated Learning (FL) and Recursive Self-Distillation (RSD) for malware detection in the Internet of Vehicles (IoV). The proposed FL-RSD framework enhances model generalization, mitigates overfitting to non-IID data, and improves adaptability to new malware variants. The RSD process iteratively transfers knowledge from a teacher model to a lightweight student model, reducing model complexity and communication overhead while preserving detection accuracy. Experimental results have confirmed that FL-RSD achieves significant performance improvements over baseline models in terms of malware detection accuracy and adversarial robustness. FL-RSD attains a malware detection accuracy and an average adversarial robustness scores over 92%, outperforming Federated Proximal, FedNova, and Hierarchical FL. The improvements range from 3% to 48% for detection accuracy, and 6% to 81% for the adversarial robustness score. Additionally, FL-RSD demonstrates a minimal memory footprint with a final global model size of 373.17 KB and maintains knowledge retention with an Average Catastrophic Forgetting Score of 93.66%. These results confirm that FL-RSD offers a lightweight, efficient, and scalable solution for malware detection in IoV.

Despite the availability of hotel booking platforms with customer reviews, selecting the best hotel or accommodation can still be difficult. This is partly because it is hard to directly contrast multiple shortlisted hotels. We propose a new visual interface, called ViSCitR, that goes beyond existing solutions in visually comparing ratings and personalizing the evaluation based on individual priorities. Targeted at a broad audience, it integrates several dimensions of comparison into a highly interactive document. The document contrasts the hotels visually and textually from different perspectives, including (i) a geographic perspective with points of interest, (ii) a comparison of ratings across different categories, (iii) rating changes over time, and (iv) summaries of the most noted positive and negative points from the review texts. The interface is accompanied by a sidebar for personalization, which supports selecting relevant points of interest and setting priorities. Changing these adapts the visual comparison across all interface sections to match the user’s preferences. A qualitative evaluation with potential hotel customers showed that the personalized visual comparison is valuable for all users, while the advanced features of the interface are leveraged to different extents.

In the dynamic field of the Industrial Internet of Things (IIoT), the networks are increasingly vulnerable to a diverse range of cyberattacks. This vulnerability necessitates the development of advanced intrusion detection systems (IDSs). Addressing this need, our research contributes to the existing cybersecurity literature by introducing an optimized Intrusion Detection System based on Deep Transfer Learning (DTL), specifically tailored for heterogeneous IIoT networks. Our framework employs a tri-layer architectural approach that synergistically integrates Convolutional Neural Networks (CNNs), Genetic Algorithms (GA), and bootstrap aggregation ensemble techniques. The methodology is executed in three critical stages: First, we convert a state-of-the-art cybersecurity dataset, Edge_IIoTset, into image data, thereby facilitating CNN-based analytics. Second, GA is utilized to fine-tune the hyperparameters of each base learning model, enhancing the model’s adaptability and performance. Finally, the outputs of the top-performing models are amalgamated using ensemble techniques, bolstering the robustness of the IDS. Through rigorous evaluation protocols, our framework demonstrated exceptional performance, reliably achieving a 100% attack detection accuracy rate. This result establishes our framework as highly effective against 14 distinct types of cyberattacks. The findings bear significant implications for the ongoing development of secure, efficient, and adaptive IDS solutions in the complex landscape of IIoT networks.

Medical imaging is an essential tool for diagnosing various healthcare diseases and conditions. However, analyzing medical images is a complex and time-consuming task that requires expertise and experience. This article aims to design a decision support system to assist healthcare providers and patients in making decisions about diagnosing, treating, and managing health conditions. The proposed architecture contains three stages: 1) data collection and labeling, 2) model training, and 3) diagnosis report generation. The key idea is to train a deep learning model on a medical image dataset to extract four types of information: the type of image scan, the body part, the test image, and the results. This information is then fed into ChatGPT to generate automatic diagnostic reports. The proposed system has the potential to enhance decision-making, reduce costs, and improve the capabilities of healthcare providers. The efficacy of the proposed system is analyzed by conducting extensive experiments on a large medical image dataset. The experimental outcomes exhibited promising performance for automatic diagnosis through medical images.

The Internet of Things (IoT) devices, networks, and applications have become an integral part of modern societies. Despite their social, economic, and industrial benefits, these devices and networks are frequently targeted by cybercriminals. Hence, IoT applications and networks demand lightweight, fast, and flexible security solutions to overcome these challenges. In this regard, artificial-intelligence-based solutions with Big Data analytics can produce promising results in the field of cybersecurity. This article proposes a lightweight dense random neural network (DnRaNN) for intrusion detection in the IoT. The proposed scheme is well suited for implementation in resource-constrained IoT networks due to its inherent improved generalization capabilities and distributed nature. The suggested model was evaluated by conducting extensive experiments on a new generation IoT security dataset ToN_IoT. All the experiments were conducted under different hyperparameters and the efficiency of the proposed DnRaNN was evaluated through multiple performance metrics. The findings of the proposed study provide recommendations and insights in binary class and multiclass scenarios. The proposed DnRaNN model attained attack detection accuracy of 99.14% and 99.05% for binary class and multiclass classifications, respectively.

Charts go hand in hand with text to communicate complex data and are widely adopted in news articles, online blogs, and academic papers. They provide graphical summaries of the data, while text explains the message and context. However, synthesizing information across text and charts is difficult; it requires readers to frequently shift their attention. We investigated ways to support the tight coupling of text and charts in data documents. To understand their interplay, we analyzed the design space of chart-text references through news articles and scientific papers. Informed by the analysis, we developed a mixed-initiative interface enabling users to construct interactive references between text and charts. It leverages natural language processing to automatically suggest references as well as allows users to manually construct other references effortlessly. A user study complemented with algorithmic evaluation of the system suggests that the interface provides an effective way to compose interactive data documents.

Abstract Data‐driven stories comprise of visualizations and a textual narrative. The two representations coexist and complement each other. Although existing research has explored the design strategies and structure of such stories, it remains an open research question how the two representations play together on a detailed level and how they are linked with each other. In this paper, we aim at understanding the fine‐grained interplay of text and visualizations in geographic data‐driven stories. We focus on geographic content as it often includes complex spatiotemporal data presented as versatile visualizations and rich textual descriptions. We conduct a qualitative empirical study on 22 stories collected from a variety of news media outlets; 10 of the stories report the COVID‐19 pandemic, the others cover diverse topics. We investigate the role of every sentence and visualization within the narrative to reveal how they reference each other and interact. Moreover, we explore the positioning and sequence of various parts of the narrative to find patterns that further consolidate the stories. Drawing from the findings, we discuss study implications with respect to best practices and possibilities to automate the report generation.

Good code quality is a prerequisite for efficiently developing maintainable software. In this paper, we present a novel approach to generate exploranative (explanatory and exploratory) data-driven documents that report code quality in an interactive, exploratory environment. We employ a template-based natural language generation method to create textual explanations about the code quality, dependent on data from software metrics. The interactive document is enriched by different kinds of visualization, including parallel coordinates plots and scatterplots for data exploration and graphics embedded into text. We devise an interaction model that allows users to explore code quality with consistent linking between text and visualizations; through integrated explanatory text, users are taught background knowledge about code quality aspects. Our approach to interactive documents was developed in a design study process that included software engineering and visual analytics experts. Although the solution is specific to the software engineering scenario, we discuss how the concept could generalize to multivariate data and report lessons learned in a broader scope.

Publication records and collaboration networks are important for assessing the expertise and experience of researchers. Existing digital libraries show the raw publication lists in author profiles, whereas visualization techniques focus on specific subproblems. Instead, we look at publication records from various perspectives mixing low-level publication data with high-level abstractions and background information. This work presents VIS Author Profiles, a novel approach to generate integrated textual and visual descriptions to highlight patterns in publication records. We leverage template-based natural language generation to summarize notable publication statistics, evolution of research topics, and collaboration relationships. Seamlessly integrated visualizations augment the textual description and are interactively connected with each other and the text. The underlying publication data and detailed explanations of the analysis are available on demand. We compare our approach to existing systems by taking into account information needs of users and demonstrate its usefulness in two realistic application examples.

Bivariate map visualizations use different encodings to visualize two variables but comparison across multiple encodings is challenging. Compared to a univariate visualization, it is significantly harder to read regional differences and spot geographical outliers. Especially targeting inexperienced users of visualizations, we advocate the use of natural language text for augmenting map visualizations and understanding the relationship between two geo-statistical variables. We propose an approach that selects interesting findings from data analysis, generates a respective text and visualization, and integrates both into a single document. The generated reports interactively link the visualization with the textual narrative. Users can get additional explanations and have the ability to compare different regions. The text generation process is flexible and adapts to various geographical and contextual settings based on small sets of parameters. We showcase this flexibility through a number of application examples.

Industrial Internet of Things (IIoT) is the other name of industrial Internet. It integrates a variety of existing industrial automation technologies with computing, machine learning, and communication technologies. Vehicular ad hoc network, an application of IIoT, is a self‐organized network of vehicles which tends to provide improved road safety, diminished traffic congestion, and ultimate comfort to the travellers. In VANETs, vehicles exchange data with each other directly or through roadside units (RSUs). Data dissemination in VANETs experiences numerous challenging issues including broadcast storm, network partitions, intermittent connectivity between vehicles, and limited bandwidth. In literature, various data dissemination schemes are proposed. However, most of these schemes are designed for either urban or highway VANET scenarios and evaluated under sparse or dense traffic conditions. Moreover, these schemes do not effectively overcome the aforementioned issues simultaneously. In this paper, we present a new data dissemination protocol for VANETs, which disseminates the emergency messages in different scenarios under varying traffic conditions. During dense traffic conditions, DDP4V employs the segmentation of transmission region of a vehicle in order to select the most appropriate next forwarding vehicle (NFV). Accordingly, it divides the transmission region of a vehicle in three distinct segments and selects vehicle(s) inside the highest priority segment to forward the message to all neighbour vehicles, whereas it also uses implicit acknowledgements for guaranteed message delivery during sparse traffic Conditions. Simulation results show that DDP4V protocol outperforms the other existing related protocols in terms of coverage, network overhead, collision, and end‐to‐end delay.