The rapid proliferation of Internet of Things (IoT) devices has increased the attack surface for cyber threats. Traditional intrusion detection systems often struggle to keep pace with novel or evolving threats. This study proposes an end-to-end generative AI-based intrusion detection and response pipeline designed for automated threat mitigation in smart home IoT environments. It leverages a Variational Autoencoder (VAE) trained on benign traffic to flag anomalies, a fine-tuned Bidirectional Encoder Representations from Transformers (BERT) model to classify anomalies into five attack categories (C&C, DDoS, Okiru, PortScan, and benign), and Grok3—a large language model—to generate tailored countermeasure recommendations. Using the Aposemat IoT-23 dataset, the VAE model achieves a recall of 0.999 and a precision of 0.961 for anomaly detection. The BERT model achieves an overall accuracy of 99.90% with per-class F1 scores exceeding 0.99. End-to-end prototype simulation involving 10,000 network traffic samples demonstrate a 98% accuracy in identifying cyber attacks and generating countermeasures to mitigate them. The pipeline integrates generative models for improved detection and automated security policy formulation in IoT settings, enhancing detection and enabling quicker and actionable security responses to mitigate cyber threats targeting smart home environments.

A number of challenging aspects have to be considered, when the Internet of Things (IoT) and Artificial Intelligence (AI) are combined into intelligent IoT systems. A key aspect that demands high attention is trustworthiness. As part of the investigations we conduct in this area in collaboration with partner companies, the need of a holistic view for trustworthiness in Intelligent IoT systems has emerged. To address such need, and to identify suitable support for it, we analyzed existing ISO standards and literature and we found out that they lack a holistic view for trustworthiness in intelligent IoT systems.To bridge this gap, we propose a conceptual model for trustworthiness in intelligent IoT systems that includes stakeholders, systems, and primary concerns, and is built upon existing standards and literature. Our model can support the design, development, operations, evolution of and communication about intelligent IoT systems. We received positive confirmation of the validity of the conceptual model from industrial practitioners working in four companies in the intelligent IoT systems area. Together with our partner companies, we plan to develop and operate approaches leveraging the conceptual model as next step.

With the increasing availability and affordability of building-integrated Heat Pumps (HPs), the number of heat pumps installed in residential buildings has risen significantly in recent years. When coupled with conventional District Heating (DH) systems in a hybrid setting, HPs provide higher energy reliability and cost-effective solutions for domestic heating. The operation of such systems, however, requires a sophisticated control system that simultaneously considers the dynamics of energy pricing and building energy needs. In this paper, we propose a nonlinear economic model predictive control to determine the optimal share for a hybrid DH-HP heating system. A resistor-capacitor thermal building model is utilized to capture the system dynamics. The results indicate that the proposed controller in the hybrid DH-HP system has a cost saving between 29% and 57% compared to the baseline scenario.

The integration of Artificial Intelligence (AI) into Internet of Things (IoT) systems has garneredconsiderable attention for its ability to enhance efficiency, functionality, and decision making.To drive further research and practical applications, it is essential to gain a deeper understandingof the different roles of AI in IoT systems. In this systematic literature review, we analyze103 articles describing Artificial Intelligence of Things (AIoT) systems found in three databases,i.e. Scopus, IEEE Xplore, and Web of Science. For each article, we examined the tasks for whichAI was used, the input and output data, the application domain, the maturity level of the system,the AI methods used, and where the AI components were deployed. As a result, we identified sixgeneral tasks of AI in IoT systems, and thirteen subtasks, the most frequent being prediction,object and event recognition, and operational decision-making. Moreover, we conclude thatmost AI components in IoT systems process numeric data as input and that healthcare isthe most common application domain followed by farming and transportation. Our analysisfurther revealed that most AIoT systems are in early development stages not validated in realenvironments. We also identified that Convolutional Neural Networks is the most frequentlyemployed AI method, with supervised learning being the dominant approach. Additionally, wefound that both AI deployment, either in the cloud or at the edge, are frequent, but that hybriddeployment is not that common. Finally, we identified key gaps in current AIoT research andbased on this, we suggest directions for future research.

This paper explores the combination of Agent-Based Social Simulation (ABSS) models. Model combination facilitates the efficient development of more complex models through reuse, enabling a more comprehensive understanding of phenomena and outcomes that individual models cannot provide on their own. Through a narrative literature review of model combination in other simulation paradigms, six different approaches were identified: Ensemble Techniques, Meta Analysis, Model Merging, Models as Modules, Model Integration and Model Chains. For each approach, examples and relevant literature is presented and current challenges are identified. Through this, the paper aims to both provide inspiration to modelers and to identify paths for future research for the combination of ABSS models and model results.

Context is an important factor in computer vision as it offers valuable information to clarify and analyze visual data. Utilizing the contextual information inherent in an image or a video can improve the precision and effectiveness of object detectors. For example, where recognizing an isolated object might be challenging, context information can improve comprehension of the scene. This study explores the impact of various context-based approaches to object detection. Initially, we investigate the role of context in object detection and survey it from several perspectives. We then review and discuss the most recent context-based object detection approaches and compare them. Finally, we conclude by addressing research questions and identifying gaps for further studies. More than 265 publications are included in this survey, covering different aspects of context in different categories of object detection, including general object detection, video object detection, small object detection, camouflaged object detection, zero-shot, one-shot, and few-shot object detection. This literature review presents a comprehensive overview of the latest advancements in context-based object detection, providing valuable contributions such as a thorough understanding of contextual information and effective methods for integrating various context types into object detection, thus benefiting researchers.

As the proliferation of Internet of Things (IoT) devices grows, challenges in security, privacy, and interoperability become increasingly significant. IoT devices often have resource constraints, such as limited computational power, energy efficiency, bandwidth, and storage, making it difficult to implement advanced security measures. Additionally, the diversity of IoT devices creates vulnerabilities and threats that attackers can exploit, including spoofing, routing, man-in-the-middle, and denial-of-service. To address these evolving threats, Intrusion Detection Systems (IDSs) have become a vital solution. IDS actively monitors network traffic, analyzing incoming and outgoing data to detect potential security breaches, ensuring IoT systems remain safeguarded against malicious activity. This study introduces an IDS framework that integrates ensemble learning with rule induction for enhanced model explainability. We study the performance of five ensemble algorithms (Random Forest, AdaBoost, XGBoost, LightGBM, and CatBoost) for developing effective IDS for IoT. The results show that XGBoost outperformed the other ensemble algorithms on two publicly available datasets for intrusion detection. XGBoost achieved 99.91% accuracy and 99.88% AUC-ROC on the CIC-IDS2017 dataset, as well as 98.54% accuracy and 93.06% AUC-ROC on the CICIoT2023 dataset, respectively. We integrate model explainability to provide transparent IDS system using a rule induction method. The experimental results confirm the efficacy of the proposed approach for providing a lightweight, transparent, and trustworthy IDS system that supports security analysts, end-users, and different stakeholders when making decisions regarding intrusion and non-intrusion events.

Given the environmental impact and cost-efficiency challenges of the conventional central District Heating (DH) systems, there is a shift towards hybrid solutions. The demand for small-scale Heat Pumps (HPs), integral components of these systems, has surged due to their electrically driven, cost-effective operation, and potential to meet environmental goals. This paper conducts a systematic literature review by investigating and highlighting hybrid heating solutions and their role in decarbonizing the built environment. It compares and discusses the potential benefits and challenges of various hybrid HP-DH systems against conventional DH-only heating approaches. The study evaluates these systems based on economic, environmental, and energy efficiency aspects, and it explores the use of intelligent and AI-based algorithms. The results indicate that, from an economic perspective, the hybrid approach can potentially offer cost savings over the long term, considering factors such as initial investment and operating expenses. The findings of the reviewed works suggest that in a DH-HP configuration, an operational cost saving between 5% and 27%, and a CO2 reduction of up to 32.3% can be achieved without additional resources. Additionally, the environmental impact analysis indicates a significant decrease in greenhouse gas emissions, aligning with global efforts to mitigate global warming.

With more hybrid heating systems available, there is a need to optimize energy use intelligently from the end-consumer perspective. This paper focuses on a multi-criteria heating system optimization to optimize cost, carbon emission, and comfort level of building occupants. A discrete Multi-Objective Model Predictive Controller (MO-MPC) algorithm is proposed to optimally utilize two heating sources connected to a building, namely district heating (DH) and a building-integrated electrical heat pump (HP). The model is tested on a real-world building case simulated with a gray box building model. The results are compared to a conventional PID controller as well as the MPC scheme, each with a single heating input, and eight different cases are constructed to make this comparison more visible. The results indicate that, using MO-MPC, a cost saving of up to 10% and emission saving of up to 13% can be reached without additional thermal discomfort, while the potential savings on cost and emission with the hybrid system can be up to 25% and 77%, respectively. Further, a sensitivity analysis on price and emission parameters is conducted to investigate the changes in the provided solution.

This study explores the use of machine learning to enhance patient mobility and caregiver ergonomics by optimizing the use of mobility aids. Traditional manual assessments can be subjective and inaccurate, so this research develops a data-driven model for object detection and human activity recognition. A computer vision dataset was created using video recordings of controlled caregiving scenarios. The study leverages advanced machine learning models, including YOLO for object detection, pose estimation, ResNet-18 for frame classification, Inception-v4 for feature extraction, and LSTM for sequence modeling. The findings provide valuable insights into integrating machine learning into mobility aids, improving both patient outcomes and caregiver well-being.