Francisco Javier Álvarez Rodríguez

La rápida expansión de la educación digital ha generado una brecha de accesibilidad, donde la evaluación manual de recursos resulta insuficiente e inviable debido al gran volumen de información disponible actualmente. Este artículo presenta una revisión sistemática de literatura con el objetivo de identificar y analizar las herramientas automáticas inteligentes existentes para la evaluación de accesibilidad digital. La metodología consistió en una revisión sistemática de literatura estructurada de diez pasos y se revisaron estudios publicados entre 2010 y 2025 en varias bases de datos; fueron seleccionados 37 artículos finales tras aplicar criterios de calidad. Los principales resultados indican un predominio de herramientas web tradicionales como WAVE, pero revelan una tendencia creciente hacia soluciones basadas en inteligencia artificial. Se concluye que, aunque la tecnología ha avanzado, existe una carencia significativa de evaluadores para documentos digitales y una necesidad urgente de herramientas que superen la validación técnica básica para comprender el contexto del contenido.

This paper presents an architectural proposal for a modular and adaptive hybrid recommender system designed to improve equitable access to PDF-based educational content for students with visual impairments. The system overcomes the limitations of traditional recommenders, which optimize rankings solely by thematic relevance while ignoring format-related barriers, by implementing a closed-loop approach that prioritizes technical accessibility prior to ranking the educational material. The architecture is structured into four independent layers: automated ingestion via multimodal artificial intelligence, dynamic student profiling, a sequential hybrid recommendation engine, and a semantic presentation interface. The technical viability of the proposal was successfully validated through a functional Python prototype integrated with the Gemini 2.5 Flash model from Google AI Studio, evaluating a repository of real-world documents split between the fields of algebra and programming. Experimental results demonstrated that the system accurately resolves the cold-start problem through an initial technical questionnaire and effectively mitigates the occurrence of false positives in practice, thanks to a reverse feedback loop that automatically updates functional security profiles and restricts resources reported with access flaws.

Recent advances in the analysis of machine learning (ML) models for accurate prediction have led to new approaches for classifying and clustering variables related to user experience (UX). In this study, however, the focus is not only on modeling variables, but also on capturing the dynamics of visual attention in relation to object properties. We evaluate models such as KNN, SVM, and K-means, which emerge as the most relevant for classifying visual attention patterns. When combined with the principles of user-centered design, this approach enables the construction of user profiles based on their interaction capabilities with the interface. Furthermore, it examines how 3D object properties—such as shape, color, shading, direction of motion, and motion acceleration—influence the information conveyed to users, thus shaping their UX directly. This method aims to enhance UX design and enable more precise representation of target users. In 11 exploration tests, the three ML models analyzed obtained an accuracy of 44, 72, and 65%.

The profile of children with ASD have difficulty interacting with conventional HCI methods. This work proposes to evaluate through eye-tracking the use of a UX focused on this profile through the design and development of mechanisms that use augmented reality and tangible interfaces. After an experimental deployment with a population of 7 children with different levels of support needs, it was possible to detect patterns among the children who had problems solving the activities. It is expected that these findings will allow a more schematic design using new technologies that will open the range of interaction methods.

This article highlights the lack of accessible educational applications for people with visual impairments and the need to integrate inclusive design from the start of development. It describes the creation of a mobile app that supports English learning for blind users in Aguascalientes, México. The project applies accessibility and usability principles, adapting functions to user capabilities and focusing on a positive experience. A key component is the MEUXuc methodology (User Experience Evaluation Methodology for Blind Users), which guides evaluation to ensure the tool meets the needs of visually impaired learners. Although they are a minority in some regions, this population is significant in absolute numbers, with many in education or professional training. The article outlines objectives such as accessible interfaces, adapted challenges, and inclusive design, offering a methodological reference for future accessible educational solutions.

La realidad aumentada (RA) es una tecnología emergente con aplicaciones significativas en la educación. Esta revisión sistemática tiene como objetivo analizar la influencia de la RA en la autoeficacia de los estudiantes, que es la creencia que tiene sobre su capacidad para organizar y ejecutar acciones dirigidas al logro de objetivos específicos (Bandura, 1977; Pajares, 1996; Gist y Mitchell, 1992) en contextos STEM (Ciencia, Tecnología, Ingeniería, Arte y Matemáticas), sintetizando la evidencia disponible entre 2018 y 2024. El estudio se desarrolló con un enfoque cualitativo y tipo documental, empleando la metodología de Barbosa et al. (2013) y utilizando bases de datos académicas como ProQuest, PubMed, Google Académico y ResearchGate. Los criterios de inclusión abarcaron: (1) estudios empíricos que exploran el uso de la RA en educación STEM, (2) investigaciones publicadas en español o inglés, y (3) análisis de impacto en la autoeficacia de los estudiantes. Para organizar la información, los estudios se agruparon según el nivel educativo, el enfoque metodológico, la región, la población objetivo y los principales resultados. Se identificaron 39 estudios que cumplieron con los criterios de inclusión, la mayoría enfocados en educación superior y secundaria, y con predominancia de métodos cuantitativos. Los hallazgos destacan que la RA fomenta la motivación y la percepción positiva del aprendizaje, especialmente en actividades prácticas. Sin embargo, los beneficios de la RA dependen del tipo de actividad, el nivel de experiencia previa de los estudiantes y la infraestructura disponible. Se recomienda ampliar el análisis a niveles educativos adicionales e incluir perspectivas de docentes para una comprensión integral del impacto de la RA en la educación.

The advancement of information and communication technologies has facilitated access to learning resources, thereby transforming the manner in which knowledge is acquired both within the context of formal education and within the context of job training. However, one of the problems that accompanies the learning resources is the lack of methods to assess their quality so that users may select the best ones in order to achieve the learning objectives. One of the attributes that has been extensively mentioned in the literature as important to achieve learning objectives is completeness, which refers to how much information a resource must contain in order to support a learning objective. However, there are not recognized methods to measure completeness. This work objective is to develop a method and a system for evaluating completeness of learning resources employed in the context of workplace training. It is crucial to ensure the completeness of a resource in order to determine whether it contains the required information to achieve the learning objective for an organization. The method for completeness evaluation comprises the formal definition of a business process, the identification of its products, and a comparison of its products with the content of the learning resources, using natural language processing and six similarity measures. Forty learning resources were assessed for completeness; the results show that a considerable number of the evaluated learning resources exhibit significant deficiencies in terms of completeness. The method is a first approach to measure completeness from a business process, and some limitations were identified. It assumes equal importance among all content elements in a resource; it penalizes extensive resources such as books, and depends heavily on well-documented business processes. Furthermore, completeness should be considered alongside other quality attributes to ensure a more comprehensive evaluation of learning resources.