Industry 4.0 has driven a paradigm shift in manufacturing, pushing industries to adopt innovative technologies for more efficient decision-making. A key component of this revolution is predictive maintenance, which plays a central role in this transformation by leveraging, among others methods, supervised machine learning techniques to anticipate equipment failures, optimise maintenance schedules, and enhance operational efficiency. This study presents a Systematic Literature Review (SLR) of 216 peer-reviewed papers published between 2019 and 2024, analysing the adoption of supervised machine learning techniques for predictive maintenance in various industrial domains, including manufacturing, machinery, energy, and smart systems. Unlike previous SLRs that broadly examine ML applications, this review provides a structured taxonomy of predictive maintenance methods, highlighting their domain-specific usage and prevalence in safety-critical industries. Additionally, we analyse the types of datasets used, revealing a strong preference for real-world data but limited public availability, which poses challenges for reproducibility and benchmarking. This study identifies key trends in ML adoption and offers insights into future research directions, thereby reinforcing the need for open datasets, explainable AI, and cross-domain generalisation.

With digital technology increasingly shaping the tourism industry, understanding customer sentiment and identifying key themes in reviews is crucial for enhancing service quality. However, traditional sentiment analysis and keyword extraction models typically demand significant time, computational resources, and labelled data for training. In this paper, we explore how Large Language Models (LLMs) can be leveraged to automatically classify reviews as positive or negative and extract relevant keywords without the need for dedicated training. Additionally, we frame the keyword extraction task as a tool to assist human users in comprehending and interpreting review content, especially in scenarios where ground truth labels for keywords are unavailable. To evaluate our approach, we conduct an experimental analysis using several datasets of tourism reviews and various LLMs. Our results demonstrate the reliability of LLMs as zero-shot classifiers for sentiment analysis and showcase the efficacy of the approach in extracting meaningful keywords from reviews, providing valuable insights into customer sentiments and preferences. Overall, this research contributes to the intersection of information technology and tourism by presenting a practical solution for sentiment analysis and keyword extraction in tourism reviews, leveraging the capabilities of LLMs as versatile tools for enhancing decision-making processes in the tourism industry.

Interest in machine learning and neural networks has increased significantly in recent years. However, their applications are limited in safety-critical domains due to the lack of formal guarantees on their reliability and behavior. This paper shows recent advances in satisfiability modulo theory solvers used in the context of the verification of neural networks with piece-wise linear and transcendental activation functions. An experimental analysis is conducted using neural networks trained on a real-world predictive maintenance dataset. This study contributes to the research on enhancing the safety and reliability of neural networks through formal verification, enabling their deployment in safety-critical domains.

Using semantic web technologies is becoming an efficient way to overcome metadata storage and data integration problems in digital archives, thus enhancing the accuracy of the search process and leading to the retrieval of more relevant results. In this paper, the results of the implementation of the semantic layer of the Józef Piłsudski Institute of America digital archive are presented. In order to represent and integrate data about the archival collections housed by the institute, the authors developed arkivo, an ontology that accommodates the archival description of records but also provides a reference schema for publishing linked data. The authors describe the application of arkivo to the digitized archival collections of the institute, with emphasis on how these resources have been linked to external datasets in the linked data cloud. They also show the results of an experiment focused on the query answering task involving a state-of-the-art triple store system. The dataset related to the Piłsudski Institute archival collections has been made available for ontology benchmarking purposes. Request access from your librarian to read this article's full text.

Modular systems are increasingly employed in critical application domains such as healthcare, smart cities, and Industry 4.0 platforms, where dynamic integration of components poses substantial challenges in ensuring consistent and secure operation. Validating system configurations against known vulnerabilities and threats requires formal, scalable, and explainable approaches. This paper presents a knowledge-driven framework designed to support validation of modular systems with respect to cyber-security threats and consistency constraints. Our approach leverages domain-specific ontologies built from well-established threat and vulnerability taxonomies and encodes inference rules to automatically detect potential threats and recommend mitigation strategies. The framework, which includes a reasoning engine and a user-friendly graphical interface providing transparent and traceable explanations for each identified threat, it is applied to a modular platform for privacy-preserving, decentralized processing of health data across European institutions. While this composable architecture enables multiple stakeholders to develop, deploy, and maintain specialised components fostering scalability and flexibility, it also introduces critical risks related to architectural coherence and security enforcement. In this context, our framework ensures a human-interpretable assessment of the system’s security posture, even in the presence of heterogeneous technologies and policies.

Cyber-Physical Systems (CPSs) are engineered systems that are built from, and depend upon, the seamless integration of computational algorithms and physical components. CPSs are widely used in many safety-critical domains, making it crucial to ensure that they operate safely without causing harm to people and the environment. Therefore, their design should be robust enough to deal with unexpected conditions and flexible to answer to the high scalability and complexity of systems. Nowadays, it is well-established that formal verification has a great potential in reinforcing safety of critical systems, but nevertheless its application in the development of industrial products may still be a challenging activity. In this paper, we describe an approach based on Satisfiability Modulo Theories (SMT) to formally verify, at the design stage, the consistency of the system design - expressed in a given domain-specific language, called QRML, which is specifically designed for CPSs - with respect to some given property constraints, with the purpose to reduce inconsistencies during the system development process. To this end, we propose an SMT-based approach for checking the consistency of configuration based-components specifications and we report the results of the experimental analysis using three different state-of-the-art SMT solvers. The main goal of the experimental analysis is to test the scalability of the selected SMT solvers and thus to determine which SMT solver is the best in checking the satisfiability of the properties.