Efficient traffic analysis and management are crucial for supporting stakeholders involved in urban and suburban road network monitoring, anomaly detection, and strategic planning. The scale and complexity of traffic data make manual approaches infeasible, necessitating automated, interpretable solutions.This study introduces TrAnSIT, an AI-driven framework that combines real-time monitoring, predictive modelling, and anomaly detection tailored to traffic management. Designed with a strong emphasis on interpretability, TrAnSIT provides clear, explainable outputs, enabling stakeholders to derive insights from complex data.Through a series of use cases, TrAnSIT is shown to enhance traffic management by identifying emerging issues, focusing on high-priority areas, and providing data-driven strategies for optimizing traffic flow and improving safety. The framework’s outputs are interpretable and tailored to various user roles, supporting both technical and non-technical decision-makers.TrAnSIT empowers stakeholders with actionable intelligence for proactive and informed decision-making in traffic management, demonstrating the value of interpretable AI frameworks in complex, data-intensive environments.

Argumentation is the inferential strategy aimed at coping with inconsistent knowledge, e.g. in order to justify or dismiss contrasting positions in a discussion. Abstract Argumentation, specifically, is based on a graph model where nodes are claims and arcs denote relationships between pair of claims. In some extensions of the original, basic framework, relationships may denote attacks or supports, and may be weighted. A characterizing feature of this setting is how the weights are composed along chains of arcs, so that the resulting indirect effect, and the final outcome of the argumentation, are intuitive and effective. Previous approaches adopted the same composition strategy for both attacks and supports. However, their meaning is different, and should result in different composition behaviors. In this paper we propose a new approach that distinguishes the two cases. We motivate it, provide computational procedures and show sample use cases demonstrating its effect on practical situations.

The growing demand for Cultural Heritage fruition is making it a major economic driver, also in connection with its tourism-related aspects. The current solutions available on the Web are still unable to provide satisfactory support to the various kinds of stakeholders and to the different applications. The History of Computing, as a peculiar, relevant and currently under investigated branch of Cultural Heritage, raises additional challenges and provides new opportunities, also in connection with a significant economic flow it generates. The variety and complexity of issues connected to this domain call for even more advanced solutions. In this paper we introduce the CHIPS&BITS project, which tackles these problems using a knowledge-based approach and leverages a novel framework that can meet the needs associated to this specific domain in a better way compared to standard Semantic Web approaches. We describe the contributions of the project to overcome the limitations of the state of the art, and focus on some of its more peculiar and original features, among which the definition of suitable ontologies to describe the complexity of the domain and the use of Artificial Intelligence algorithms for giving the users an advanced and personalized experience.

This research introduces an artificial intelligence-based strategy for improving symbol recognition within the field of library science, concentrating on the creation and application of sophisticated technological solutions. Consistent with the objectives of the CHANGES project—Cultural Heritage Active Innovation for Sustainable Society, which focuses on the enhancement and management of cultural heritage through a multidisciplinary and interinstitutional approach—this strategy employs convolutional neural networks (CNNs) for accurate symbol classification. A CNN model was developed using an extensive dataset comprising over 6000 symbols, implementing meticulous preprocessing, feature extraction, and supervised learning protocols. The methodological pipeline incorporates advanced image segmentation techniques to isolate symbols from complex manuscripts, followed by data augmentation to enhance model resilience. The system is supported by a high-performance computing framework to manage large datasets efficiently, thereby facilitating more precise identification and analysis. This integration of machine learning techniques, exhaustive data management, and computational capabilities significantly advances existing symbol recognition methodologies, providing scholars with a potent tool for assisting in the classification and interpretation of historical symbols. The findings corroborate the potential of AI-enhanced symbol recognition in contributing to the broader objectives of computational library science and historical research.

Cultural Heritage is opening up from the professional community to a wider public, generating an increasing demand for culture and an associated economic turnaround. This step requires to differentiate the behavior of Cultural Heritage systems, dealing with a wide variety of backgrounds, expectations, contexts, aims, educational and cultural level, preferences and interests. Computer Science and Artificial Intelligence can play a key role in this landscape, finetuning the fruition of cultural items to every kind of stakeholder and even to single users. In this paper we present an approach to personalization of Cultural Heritage fruition based on Knowledge Graphs. An approach to describe user models, and to use them for extracting personalized information, is proposed, and a platform that embeds this approach is described.

This study proposes a novel, scalable framework for the automated classification and synthesis of survey literature by integrating state-of-the-art Large Language Models (LLMs) with robust ensemble voting techniques. The framework consolidates predictions from three independent models—GPT-4, LLaMA 3.3, and Claude 3—to generate consensus-based classifications, thereby enhancing reliability and mitigating individual model biases. We demonstrate the generalizability of our approach through comprehensive evaluation on two distinct domains: Question Answering (QA) systems and Computer Vision (CV) survey literature, using a dataset of 1154 real papers extracted from arXiv. Comprehensive visual evaluation tools, including distribution charts, heatmaps, confusion matrices, and statistical validation metrics, are employed to rigorously assess model performance and inter-model agreement. The framework incorporates advanced statistical measures, including k-fold cross-validation, Fleiss’ kappa for inter-rater reliability, and chi-square tests for independence to validate classification robustness. Extensive experimental evaluations demonstrate that this ensemble approach achieves superior performance compared to individual models, with accuracy improvements of 10.0% over the best single model on QA literature and 10.9% on CV literature. Furthermore, comprehensive cost–benefit analysis reveals that our automated approach reduces manual literature synthesis time by 95% while maintaining high classification accuracy (F1-score: 0.89 for QA, 0.87 for CV), making it a practical solution for large-scale literature analysis. The methodology effectively uncovers emerging research trends and persistent challenges across domains, providing researchers with powerful tools for continuous literature monitoring and informed decision-making in rapidly evolving scientific fields.

Non-communicable diseases (NCDs) like hypertension, diabetes, osteoporosis, and cancer constitute 80

Sub-symbolic Machine Learning (ML) techniques, and specifically Neural Network-based ones, recently took over the research landscape, thanks to their efficiency and impressive effectiveness. On the other hand, the recent debate on ethics and AI and the first regulations on AI are progressively calling for anthropocentricity, which in turn requires explicit, human-understandable, and explainable approaches and representations that allow humans to be active parts in the loop. In these cases, logic-based approaches are more suitable. The Inductive Logic Programming (ILP) branch of research in ML provides an anwer to this need and a uniform and unifying framework for three relevant industrial and research concerns: management of databases, implementation of software systems, and modeling of human-like reasoning strategies. A particular ILP framework based on the Object Identity (OI) assumption was proposed in the 1990s, for which desirable theoretical and pratical properties were demonstrated and working tools and systems that successfully approached real-world and classical problems in AI were developed. In an age when mainstream research and media seem to reduce AI and ML to just deep learning, this paper celebrates the 30th anniversary of OI by providing for the first time a comprehensive overview of the framework to be used as a reference for researchers still interested in investigating the ILP approach to ML.

The increasing integration of Artificial Intelligence (AI) in education has led to the development of innovative tools like Intelligent Question-Answering Systems (IQASs), aiming to revolutionize traditional learning paradigms. However, many existing IQAS struggle with the nuances of natural language and the complexities of student questions. This research focuses on developing a context-aware IQAS that leverages advanced Natural Language Processing (NLP) techniques and contextual information, including student learning history and educational content, to provide personalised support. This study also introduces a software tool that utilizes NLP techniques to automatically generate FAQs from educational materials. Employing a hybrid approach combining rule-based and machine learning techniques, the IQAS demonstrated high accuracy in interpreting and responding to a wide range of student queries. The software tool effectively automated the generation of FAQs, creating a valuable resource for personalised learning. The findings suggest that these tools can significantly improve student engagement, motivation, and learning outcomes, highlighting the potential of AI to transform education and pave the way for more personalised, adaptive, and effective learning environments.

Digital Libraries, especially large ones, need advanced tools for analyzing and interpreting their content. This paper introduces a novel Web-based tool designed to support library users in browsing and examining the content of digital libraries, and in extracting and interpreting high-level information from them. Leveraging external data and metadata repositories, our software uses state-of-the-art solutions in Natural Language Processing and Visualization to analyze and identify thematic trends and clusters over time. We describe the tool's architecture and functionality, showing its effectiveness in tracking the evolution of topics along time within a textual corpus. We further illustrate it through a case study that analyzes two decades of scholarly articles from an established conference on Digital Libraries, showcasing the tool's potential to help in understanding past and current academic research, and to guide future trajectories in this field.

Ontologies are essential for the management and integration of heterogeneous datasets. This paper presents OntoBuilder, an advanced tool that leverages the structural capabilities of semantic labeled property graphs (SLPGs) in strict alignment with semantic web standards to create a sophisticated framework for data management. We detail OntoBuilder’s architecture, core functionalities, and application scenarios, demonstrating its proficiency and adaptability in addressing complex ontological challenges. Our empirical assessment highlights OntoBuilder’s strengths in enabling seamless visualization, automated ontology generation, and robust semantic integration, thereby significantly enhancing user workflows and data management capabilities. The performance of the linked data tools across multiple metrics further underscores the effectiveness of OntoBuilder.

The traditional record-based approach to the description of Cultural Heritage is nowadays obsolete. It is unable to properly handle complex descriptions and it cannot support advanced functions provided by Artificial Intelligence techniques for helping practitioners, scholars, researchers and end-users in carrying out their tasks. A graph-based, semantic approach is needed, such as that provided by Semantic Web solutions. Also, a ‘holistic’ description approach is needed, that includes and inter-connects all branches and types of Cultural Heritage, and that is not limited to describing just the formal metadata of cultural objects, but can deal with their content, physicality, context and lifecycle, as well. The GraphBRAIN framework and technology for Knowledge Graph management enforces all these ideas and enjoys improved efficiency, expressiveness, and flexibility thanks to the use of the LPG model for knowledge representation. This paper describes GraphBRAIN and its application to several Cultural Heritage-related fields, including digital libraries, archives and museums, history of computing, and tourism as a way to boost fruition of these items.