ABSTRACT This article aims to analyze the level of research maturity on cybersecurity in the Metaverse through the identification and synthesis of the technological, organizational, human, and regulatory factors that influence Metaverse cybersecurity. To this end, literature was collected from Web of Science and Scopus, applying PRISMA 2020, bibliometric techniques, and “scientific mapping” to analyze 285 documents retrieved from both databases. By using tools such as SciMap and VOSViewer, it was possible to develop a knowledge diagram that illustrates the findings. The results reveal seven clusters organized around major factor domains: (1) Internet of Things and Machine Learning; (2) The Metaverse ecosystem; (3) Hardware infrastructure; (4) Secure data processing and access control; (5) Blockchain technology; (6) Authentication; (7) Artificial Intelligence, augmented reality, and quantum technologies. The conclusions indicate that human and governance factors receive comparatively less attention than technological-algorithmic ones, opening prioritized lines of inquiry for future research and public policy.

This study evaluates the classification performance of four GPT-based models (GPT-4.1, GPT-4.1-mini, GPT-4.1-nano, and o4-mini) under zero-shot prompting conditions on the complete, multilingual CoDA dataset of Dark Web content, comprising 10 illicit activity categories. The models GPT-4.1, GPT-4.1-mini, and o4-mini achieve a weighted F1 score of 0.885, surpassing prior zero-shot baselines on this dataset. Stability analysis using TARa@10 demonstrates high output consistency for GPT-4.1 (0.964) and GPT-4.1-mini (0.970), indicating their reliability for operational use. Multilingual evaluation reveals only a modest English vs. non-English performance gap for GPT-4.1 (0.031), while other models perform comparably across languages. The strongest results appear in Drugs , Gambling , and Porn (F1 0.9), whereas lower scores are observed in ambiguous or overlapping categories like Violence (F1 0.76) or Crypto (F1 0.84). A qualitative review of misclassifications suggests that some model predictions align with reasonable semantic interpretations, potentially highlighting annotation inconsistencies. This work establishes a performance baseline for GPT-based models in zero-shot classification of multilingual Dark Web content and underscores the importance of clear category definitions for effective deployment.

This research addresses the complex challenge of automated repair of code vulnerabilities, vital for enhancing digital security in an increasingly technology-driven world. The study introduces a novel and efficient format for the representation of code modification, using advanced Large Language Models (LLMs) such as Code Llama and Mistral. These models, fine-tuned on datasets featuring C/C++ code vulnerabilities, significantly improve the accuracy and adaptability of automated code repair techniques. A key finding is the enhanced repair accuracy of these models when compared to previous methods such as VulRepair, which underscores their practical utility and efficiency. The research also offers a critical assessment of current evaluation metrics, such as “Perfect Predictions”, and their limitations in reflecting the true capabilities of automated repair models in real-world scenarios. Following this, it underscores the importance of using test datasets devoid of train samples, emphasizing the need for dataset integrity to enhance the effectiveness of LLMs in code repair tasks. The significance of this work is its contribution to digital security, setting new standards for automated code vulnerability repair and paving the way for future advancements in the fields of cybersecurity and artificial intelligence. The study does not only highlight the potential of LLMs in enhancing code security but also fosters further exploration and research in these crucial areas.

There are many datasets for training and evaluating models to detect web attacks, labeling each request as normal or attack. Web attack protection tools must provide additional information on the type of attack detected, in a clear and simple way. This paper presents a new multi-label dataset for classifying web attacks based on CAPEC classification, a new way of features extraction based on ASCII values, and the evaluation of several combinations of models and algorithms. Using a new way to extract features by computing the average of the sum of the ASCII values of each of the characters in each field that compose a web request, several combinations of algorithms (LightGBM and CatBoost) and multi-label classification models are evaluated, to provide a complete CAPEC classification of the web attacks that a system is suffering. The training and test data used for training and evaluating the models come from the new SR-BH 2020 multi-label dataset. Calculating the average of the sum of the ASCII values of the different characters that make up a web request shows its usefulness for numeric encoding and feature extraction. The new SR-BH 2020 multi-label dataset allows the training and evaluation of multi-label classification models, also allowing the CAPEC classification of the various attacks that a web system is undergoing. The combination of the two-phase model with the MultiOutputClassifier module of the scikit-learn library, together with the CatBoost algorithm shows its superiority in classifying attacks in the different criticality scenarios. Experimental results indicate that the combination of machine learning algorithms and multi-phase models leads to improved prediction of web attacks. Also, the use of a multi-label dataset is suitable for training learning models that provide information about the type of attack.

Nowadays, the Zero Trust model has become one of the standard security models. This paradigm stipulates as mandatory the protection of each endpoint, looking for providing security to all the network. To meet this end, it is necessary to guarantee the integrity of the access control systems. One possibility for bringing security to the different endpoints is continuous authentication, as an access control system. Continuous authentication is the set of technologies capable of determining if a user’s identity remains in time; whether he is the legitimate user (i.e., the only one who should know the secret credentials) or the identity has been impersonated by someone else after the authentication’s process was completed. Continuous authentication does not require the active participation of the user. Aiming to identify the different technologies involved in continuous authentication’s implementations, evaluation methods, and its use cases, this paper presents a systematic review that synthesizes the state of the art. This review is conducted to get a picture about which data sources could allow continuous authentication, in which systems it has been successfully implemented, and which are the most adequate ways to process the data. This review also identifies the defining dimensions of continuous authentication systems.

Classroom response systems (CRSs) are tools that allow all students to respond to teacher questions individually through a software platform during lectures. Some widely used CRSs, like Kahoot or Classtime, include gamified designs that transform teacher questions into competitive or cooperative challenges in an attempt to make their usage more enjoyable. A question arises as to what differences may exist in the effects of competitive and cooperative gamified CRSs on student's test performance and learning. This study aims to analyze those differences in the context of a pre-graduate programming university course, which introduces students to video game programming. We present a quasi-experiment with 69 students assigned to two experimental and one control group. We examine between-group differences in CRSs' quiz results and programming skills when using competitive, cooperative, and neutral CRSs. An initial analysis suggests that quiz results are significantly lower when using competitive CRSs, in contrast with cooperative and neutral CRSs, but that there are no significant differences in the programming skill acquisition, considering practical assignments, as well as midterm and final examinations.

Mobile devices have small screens that require the user to constantly resize the web content in order to read the text. This study sets out to (a) check whether current mobile devices are able to zoom or increase the font size of text on websites and (b) check the most efficient, effective, and satisfying usability method for the user to do so. A preparation stage with mobile devices and an experiment with real users were designed and carried out. Results suggested that current mobile devices are able to resize text in web pages. User testing returned that there is no difference in the ability to perform the task or in errors or time to complete it between text-resizing methods. However, users preferred increasing/decreasing font size links method in nontouch devices. In the light of results, it is recommended that links are provided to resize the text in web contents and that they are made readily available.

In this paper we propose a system that allows to adapt educational contents to learners, based on the knowledge of a learner who is conducting the training through his mobile device, contents will also be tailored to the features of the device and to the context where the learner is at the present moment.

Inspired by political parties' behavior in parliament's elections of chairman, Parliameantary Optimization Algorithm (POA) has emerged as a new stochastic population-based optimizer. Current research has proven POA efficiency in numerical optimization but it is difficult to find a POA version that deals with combinatorial optimization. In this paper we present a parliamentary algorithm that can solve permutation constraint satisfaction problems along with the results of its experimental testing and comparison with other evolutionary methods. Results demonstrate POA efficiency in this new landscape.

This paper presents a web tool that uses Java Server Faces to design questions graphically, to test the outcome in a learning environment and to export them using the QTI v2.1 specification in order to ensure interoperability between different learning systems.

This paper proposes to improve the Simple Query Interface (SQI), designed to query learning object repositories, modifying a few existing methods to generate new ones. Replacement is not intended, instead, our proposal is to create new methods that incorporate several enhancements in order to enable compatibility with currently compliant systems that use the old methods. Besides, new methods are also proposed aiming at improving the interface as well as at providing compatibility with the Simple Publishing Querying (SPI) specification. With these improvements SQI specification will perform its task in a more structured and efficient manner.

In this paper we present our experience in building and lecturing an interdisciplinary CS course aimed at teaching the Unified Modeling Language (UML) in a cultural context. Combining modeling concepts and ideas brought from philosophy and history of ideas we present a novel way to teach UML and modeling beyond its original environment. Course design, implementation and results are presented. Although the course was very experimental results were quite positive.

This paper presents a course that explores the borders of database systems and their relations with the philosophical movement of phenomenology, all framed in a humanistic informatics context.

This paper presents a new mobile application designed for auto-assessment that allows students to test their knowledge and expertise in a specific topic using questionnaires designed by their teachers. Students' achievement was evaluated and results suggest that this kind of approaches can improve it.

The process of creating e-learning contents using reusable learning objects (LOs) can be broken down in two sub-processes: LOs finding and LO sequencing. Sequencing is usually performed by instructors, who create courses targeting generic profiles rather than personalized materials. This paper proposes an evolutionary approach to automate this latter problem while, simultaneously, encourages reusability and interoperability by promoting standards employment. A model that enables automated curriculum sequencing is proposed. By means of interoperable competency records and LO metadata, the sequencing problem is turned into a constraint satisfaction problem. Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) agents are designed, built and tested in real and simulated scenarios. Results show both approaches succeed in all test cases, and that they handle reasonably computational complexity inherent to this problem, but PSO approach outperforms GA.

Standards and specifications widely accepted and used lay the foundations to enable and facilitate the interoperability among systems, and the software maintenance and reuse, especially within the scope of learning objectspsila search systems. One of these standards is the SQI (standard query interface) specification by the European Committee for Standardization in which many search systems, including public ones, are based. This paper analyzes the degree of compliance with this specification by a significant number of learning objects repositories.

The University of Alcalá (Spain) and the Lund University (Sweden) have been collaborating for several years in the development of various courses at the in-between of the Computer Science and the Humanities fields. We are trying to join different expertise and to apply it to design new ways of teaching. Within this scope we present here a course that is aimed to teach the Unified Modeling Language (UML) to humanities students.

In e-learning initiatives content creators are usually required to arrange a set of learning resources in order to present them in a comprehensive way to the learner. Course materials are usually divided into reusable chunks called Learning Objects (LOs) and the ordered set of LOs is called sequence, so the process is called LO sequencing. In this paper an intelligent agent that performs the LO sequencing process is presented. Metadata and competencies are used to define relations between LOs so that the sequencing problem can be characterized as a Constraint Satisfaction Problem (CSP) and artificial intelligent techniques can be used to solve it. A Particle Swarm Optimization (PSO) agent is proposed, built, tuned and tested. Results show that the agent succeeds in solving the problem and that it handles reasonably combinatorial explosion inherent to this kind of problems.

The process of creating e-learning contents using reusable learning objects (LOs) can be broken down in two sub-processes: LOs finding and LO sequencing. Sequencing is usually performed by instructors, who create courses targeting generic profiles rather than personalized materials. This paper proposes an evolutionary approach to automate this latter problem while, simultaneously, encourages reusability and interoperability by promoting standards employment. A model that enables automated curriculum sequencing is proposed. By means of interoperable competency records and LO metadata, the sequencing problem is turned into a constraint satisfaction problem. Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) agents are designed, built and tested in real and simulated scenarios. Results show both approaches succeed in all test cases, and that they handle reasonably computational complexity inherent to this problem, but PSO approach outperforms GA.

As a part of many e-learning initiatives, a set of learning units must be arranged in a particular order to meet the learners’ requirements. This process is known as sequencing and it is typically performed by instructors, who create wide-public ordered series rather than learner personalized sequences. This paper proposes an innovative intelligent technique for learning object automated sequencing using particle swarms. E-Learning standards are promoted in order to ensure interoperability. Competencies are used to define relations among learning objects within a sequence, so that the sequencing problem turns into a permutation problem and a particle swarm optimization algorithm can be applied to solve it. Results demonstrate that the new agent succeeds and it shows a good performance in real and tests scenarios.

Current e-learning applications should provide universal access to educational information, regardless of the standards, protocols or programming languages used. This would maximize the reuse of learning objects and help this technology to gain its place as a fundamental and, in some cases only, support tool for present day educational systems. To achieve this, it is necessary that standards are adopted that guarantee, both the development of learning objects and the search systems that provide access to the repositories that contain them; this being the only way to guarantee universal access.

The process of creating e-learning courseware using reusable learning objects (LOs) can be broken down in two sub-processes: LOs finding and LO sequencing. Sequencing is usually performed by instructors, who create courses targeting generic profiles rather than personalized materials. This paper proposes an evolutionary approach to automate this latter problem while, simultaneously, encourages reusability and interoperability by promoting standards employment.

In e-learning initiatives, sequencing problem concerns arranging a particular set of learning units in a suitable succession for a particular learner. Sequencing is usually performed by instructors, who create general and ordered series rather than learner personalized sequences. This paper proposes an innovative intelligent technique for learning object automated sequencing using particle swarms. E-learning standards are promoted in order to ensure interoperability. Competencies are used to define relations between learning objects within a sequence, so that the sequencing problem turns into a permutation problem and AI techniques can be used to solve it. Particle Swarm Optimization (PSO) is one of such techniques and it has proven with good performance solving a wide variety of problems. An implementation of the PSO, for learning object sequencing, is presented and its performance in a real scenario is discussed.