This paper explores the integration of Artificial Intelligence Generated Content (AIGC), a rapidly evolving branch of generative AI, with Human-Machine intelligence (HMI) to enhance the functionality of Intelligent Transportation Systems (ITS). As transportation systems grow increasingly complex, adaptive decision-making becomes essential for interpreting vast streams of real-time data from vehicles, infrastructure, and users. AIGC plays a transformative role in optimizing traffic flow through dynamic routing and real-time traffic management, while human intelligence ensures these systems remain responsive to evolving real-world conditions. For safety, AIGC is used to simulate complex driving scenarios for autonomous vehicle training and detect traffic anomalies, with human oversight providing contextual decisions in ambiguous situations. For sustainability, AIGC supports data-driven strategies to reduce emissions and energy use, while human expertise ensures alignment with ethical and environmental goals. This synergy enhances real-time decision-making, improving both accuracy and adaptability across ITS scenarios. The paper presents a comprehensive review of core and supporting AIGC technologies and their applications across key ITS domains. Case studies and initiatives from industry leaders demonstrate practical implementations of AIGC-driven HMI collaboration. To guide future deployments, we propose a conceptual five-layer evaluation framework for assessing AIGC-HMI systems, encompassing functional performance, human interaction, explainability, ethical compliance, and robustness. We also address challenges such as legacy system integration, data privacy, model bias, and scalability. The paper concludes by outlining future research directions, emphasizing the need for scalable, interpretable, and ethically aligned AIGC models. This work contributes to the development of intelligent, adaptive, and trustworthy transportation systems. 

Cyberbullying has become a major challenge in the digital era, and many people, especially adolescents, use social media platforms to communicate and share information. Some exploit these platforms to embarrass others through messages, e-mails, speech, and public posts, causing severe psychological harm to victims. This study reviews existing research on technologies, approaches, datasets, and evaluation metrics for cyberbullying detection, while highlighting future directions and key challenges. The findings show that traditional models work reasonably well with small datasets but require constant updates; machine learning models face feature extraction and linguistic limitations; deep learning models perform better but lack multilingual and cross-lingual capabilities; and large language models (LLMs) achieve the highest performance, offering flexibility and rich linguistic features but face issues of high-energy use and real-time applicability. Addressing technological, methodological, dataset, and linguistic challenges will improve cyberbullying detection, helping to protect online communication and promote social responsibility. 

Artificial Neural Networks (NNs) are unable to learn tasks continually using a single model, which leads to forgetting old knowledge, known as catastrophic forgetting. This is one of the shortcomings that usually plague intelligent systems based on NN models. Federated Learning (FL) is a decentralized approach to training machine learning models on multiple local clients without exchanging raw data. A paradigm that handles model learning in both settings, federated and continual, is known as Federated Continual Learning (FCL). In this work, we propose a novel FCL algorithm, called FedCluLearn, which uses a stream micro-cluster indexing scheme to deal with catastrophic forgetting. FedCluLearn interprets the federated training process as a stream clustering scenario. It stores statistics, similar to micro-clusters in stream clustering algorithms, about the learned concepts at the server and updates them at each training round to reflect the current local updates of the clients. FedCluLearn uses only active concepts in each training round to build the global model, meaning it temporarily forgets the knowledge that is not relevant to the current situation. In addition, the proposed algorithm is flexible in that it can consider the age of local updates to reflect the greater importance of more recent data. The proposed FCL approach has been benchmarked against three baseline algorithms by evaluating its performance in several control and real-world data experiments. The implementation of FedCluLearn and the experimental results are available at https://github.com/milenaangelova1/FedCluLearn.

Integrating artificial intelligence (AI) with immersive analytics (IA) represents a promising means of leveraging advanced computational techniques to enhance data visualization and analysis. This study examines the state-of-the-art of AI-IA integration by addressing three key research issues: the significant application domains, the AI techniques used and their combinations, and current challenges and future directions. Results of reviewing 43 relevant studies reveal that AI-IA integration is still in its early stages, as existing research has mainly focused on a limited range of data types and application scenarios. By analyzing the application domains, this systematic literature review supports previous findings of important applications in the fields of education, manufacturing, and healthcare. At the same time, it identifies emerging applications that have progressed from XR and AI domains to AI-IA integration, such as sports events, assistive systems, urban planning, and disaster management. We contribute to extending established visual analytics (VA) pipelines into XR environments with integrated AI techniques. AI techniques are identified as contributing in five ways to this IA pipeline. Our contribution also includes identifying four key challenges and seven opportunities for future exploration. The review concludes that combining AI and IA holds the potential to create innovative applications using advanced AI and immersive visualization techniques. We present an overview of these applications and address key issues for future development. 

Key Encapsulation Mechanisms (KEM) is a special case of Public Key Encryption (PKE) that was recently standardized by National Institute of Standards and Technology in USA. The broader adoption of the term in industry practice was necessitated by the discovery of the malleability property of ciphertext, which led to new approaches PKE. New standard initiated refactoring of all cryptographic software libraries and this process relates to the problems of cryptographic agility. This Systematization of Knowledge (SoK) addresses the developments of public key encryption methods and the main challenges that drive the specialization of KEM in cryptographic software. Based upon our findings from a systematic literature review, we present a formal analysis that provides cryptographic users a means of better understanding of KEM and its roles in cryptographic migrations in IoT systems. We have identified the main milestones of KEM evolution and structured it into four development areas. We found that the evolution of KEM is defined by a variety of mathematical foundations that always reflect various aspects of the cryptosystem. Our findings indicate that academia, industry practitioners and standardization bodies propagate such approaches into practice by additional abstraction layers in cryptographic software libraries. However, the libraries is still not in consensus, which is confirmed after the discovery of a new class of libraries, cryptographic bindings. To structure the mentioned phenomena, we introduced a novel, three-facet, consumer-centered mapping of the data security domain.We believe our contribution can help researchers and practitioners to have a broader and deeper understanding of data encryption tooling in context of cryptographic migrations.

Many machine learning models deployed on smart or edge devices experience a phase where there is a drop in their performance due to the arrival of data from new domains. This paper proposes a novel unsupervised domain adaptation algorithm called DIBCA++ to deal with such situations. The algorithm uses only the clusters’ mean, standard deviation, and size, which makes the proposed algorithm modest in terms of the required storage and computation. The study also presents the explainability aspect of the algorithm. DIBCA++ is compared with its predecessor, DIBCA, and its applicability and performance are studied and evaluated in two real-world scenarios. One is coping with the Global Navigation Satellite System activation problem from the smart logistics domain, while the other identifies different activities a person performs and deals with a human activity recognition task. Both scenarios involve time series data phenomena, i.e., DIBCA++ also contributes towards addressing the current gap regarding domain adaptation solutions for time series data. Based on the experimental results, DIBCA++ has improved performance compared to DIBCA. The DIBCA++ has performed better in all human activity recognition task experiments and 82.5% of experimental scenarios on the smart logistics use case. The results also showcase the need and benefit of personalizing the models using DIBCA++, along with the ability to transfer new knowledge between domains, leading to improved performance. The adapted source and target models have performed better in 70% and 80% of cases in an experimental scenario conducted on smart logistics. 

The AI marketplace ecosystem accelerates multiple modules of the AI engineering pipeline by fostering collaboration between stakeholders. However, marketplace collaborators often face a dilemma in striking a balance between sharing artifacts and protecting intellectual property (IP) rights. Thus, there is a need for a license management system within the AI marketplace to facilitate the exchange of artifacts in a trusted and secure manner. 

This work shares experiences while building such a license management system within the Bonseyes marketplace (BMP), a functional crowdsourcing AI marketplace that specializes in deploying real-time applications on edge devices. The BMP was developed, and its applicability is proven through the European H2020 project by a series of open calls and workshops, for gathering stakeholders and orchestrating the marketplace operations. 

The main contributions of this work are (i) implementation of an end-to-end license management system that deals with selecting license templates, license agreement interaction between seller and buyer, and the generation and enforcement of human- and machine-readable license files, and (ii) introduction of "Synchronization licenses'' concept from the music industry to the AI marketplace context where consumers acquire a license to integrate the artifact into another application, and a respective BMP use-case for collaborative AI engineering. 

Multi-user extended reality (XR) is an advanced technology that immerses users simultaneously into the meta-verse, aiming to provide a real-time interaction experience in different scenarios. However, different users' network conditions cause data transmission to be asynchronous. It forces users' behavior in the metaverse to go out of sync, seriously impacting the user experience. This paper presents a framework for human-centred synchronization of networks and metaverse applications, often denoted as Intelligent Realities, using Digital Twin (DT) technologies for the networks (network DTs) as well as for the individual users' XR parts (XR DTs). Our framework uses users' attention as a key input to dynamically allocate network resources and align content, thereby keeping widely distributed users in sync. To motivate this framework, we surveyed recent DT-based XR and network systems and found a significant gap: few studies work on synchronizing systems between network DTs and XR DTs. This highlights the need to explore integrated solutions for networked real-time multi-user interaction in the metaverse. 

Split Learning (SL) is an emerging decentralized paradigm that enables numerous participants, to train a deep neural network without disclosing sensitive information, such as patient data, in fields such as healthcare. In healthcare, SL enables distributed training across a variety of medical devices, hospitals, and organizations, improving model robustness while maintaining patient confidentiality. However, training models within SL is affected by data heterogeneity and sensitivity, and often requires more computational resources than an individual data provider can afford. This can result in significant model divergence and decreased performance due to differences in data distributions between various clients. To address this issue, we propose a framework that integrates fairness and adaptivity considerations, called ADF-SL. In particular, ADF-SL dynamically adjusts the total number of clients involved in model training and the number of iteration required to achieve convergence without compromising participant privacy. To evaluate performance, we compare the effectiveness of ADF-SL with that of the naive (Vanilla) SL approach, SplitFed and FairFed. Extensive experiments performed on time series electrocardiogram (ECG) databases (MITDB, SVDB, and INCARTDB) indicate that ADF-SL significantly outperforms the three existing algorithms that served as baselines. Compared to these baseline methods, ADF-SL accelerates model training on clients by up to 22.7%, 10.4%, and 5.8% compared to Vanilla SL, SplitFed, and FairFed, respectively, while maintaining model convergence and accuracy. Furthermore, the conducted ablation study has confirmed the importance of ADF-SL decay enrichment, which has outperformed non-decay ADF-SL for each used dataset by up to 15.8%, 43.9%, and 7.6%, respectively. 

In this study, we propose a novel neuro-symbolic approach to deal with the inherent ambiguity in image scene classification, combining the usage of pre-trained deep learning (DL) models with concepts from modal logic and evidence theory. The DL models are used to detect objects and estimate their depth in a set of labeled images. The obtained outputs are employed to form a dataset of instances characterizing the possible classes. Subsequently, a multi-valued mapping is defined between the data instances and the considered images resulting into each image being represented by the set of instances associated with it. The obtained mapping is utilized to infer necessity and possibility conditions of each class, or equivalently its upper (plausibility) and lower (belief) probabilities. Based on these interval evaluations, a rule-based and a score-based classifiers are built. The overall method is explainable and directly interpretable, robust to data scarcity and data imbalance. The presented framework is studied and evaluated on an abandoned bag detection use case.