Abstract Museums widely use audio guides, yet these are delivered identically to all visitors regardless of their profile. This study addresses that gap by combining GPT-4 with Falk and Dierking’s visitor categorization framework to generate personalized audio guides and examine whether LLM-generated content can effectively meet the distinct needs of five visitor types: Explorers, Experience Seekers, Professionals/Hobbyists, Facilitators, and Rechargers. Personalized and general audio guides were generated for three artworks using a prompt-chain approach encoding visitor-specific needs. A user study with 56 self-identified participants evaluated the content, complemented by expert validation assessing factual accuracy and domain-specific quality. Personalized guides outperformed general ones across most categories and needs, with notable gains in engagement for Experience Seekers (+8.9%), curiosity stimulation for Explorers (+8.8%), and work/hobby relevance for Professionals/Hobbyists (+10.1%). Preference, however, was moderated by artwork characteristics. Expert review confirmed factual soundness while identifying gaps in domain-specific terminology and tonal alignment. These findings directly motivated our Curator-AI collaborative framework, which organizes content production into four phases, with the curator as the governing intelligence. This work demonstrates the potential of LLMs to support personalization in cultural heritage contexts through established visitor categorization frameworks. It also highlights the importance of a human-in-the-loop approach, in which curators remain actively involved in supervising, validating, and refining AI-generated content. Overall, the proposed framework suggests a viable path toward scalable and personalized museum experiences that balance automation with curatorial oversight.

Indoor Positioning Systems (IPSs) hold significant potential for enhancing visitor experiences in cultural heritage. By enabling personalized navigation, efficient artifacts organization, and better interaction with exhibits, IPSs can transform how individuals engage with museums, galleries and libraries. However, these institutions face several challenges in implementing IPSs, including environmental constraints, technical limits, and limited experimentation. Received Signal Strength (RSS)-based approaches using Bluetooth Low Energy (BLE) and WiFi have emerged as preferred solutions due to their non-invasive nature and minimal infrastructure requirements. Nevertheless, the lack of publicly available RSS datasets that specifically reflect museum environments presents a substantial barrier to developing and evaluating positioning algorithms designed for the intricate spatial characteristics typical of cultural heritage sites. To address this limitation, we present BAR, a novel RSS dataset collected in front of 90 artworks across 13 museum rooms using two different platforms, i.e., Android and iOS. We provide an advanced position classification baseline taking advantage of a proximity-based method and k-NN algorithms. In our experiments, room-level accuracy ranges from 92.36 % to 99.97 % and artwork Top-3 from 75.15 % to 98.26 %, depending on the configuration, with cross-platform scenarios revealing significant challenges.

Concept maps are significant tools able to support several tasks in the educational area such as curriculum design, knowledge organization and modeling, students’ assessment and many others. They are also successfully used in learning activities in which students have to represent domain knowledge according to teacher’s assignment. In this context, the development of Learning Analytics approaches would benefit of methods that automatically compare concept maps. Detecting concept maps similarities is relevant to identify how the same concepts are used in different knowledge representations. Algorithms for comparing graphs have been extensively studied in the literature, but they do not appear appropriate for concept maps. In concept maps, concepts exposed are at least as relevant as the structure that contains them. Neglecting the semantic and didactic aspect inevitably causes inaccuracies and the consequently limited applicability in Learning Analytics approaches. In this work, starting from an algorithm which compares didactic characteristic of concept maps, we present an extension which exploits a semantic approach to catch the actual meaning of the concepts expressed in the nodes of the map.

Wikipedia is one of the most used repositories of free learning materials published by communities of experts: most of the students, and teachers do great use of it, despite the criticism about its reliability. Here we address the problem of helping teachers to build courses with Wikipedia pages to prepare on-the-fly courses. We use a web-based system called Wiki Course Builder, which allows teachers to query Wikipedia, by submitting a set of keywords, and having as a response four different ranked lists of retrieved Wikipedia pages, depending each list on a given retrieval metric implemented. In addition to the classic TF-IDF, Information Gain and Latent Semantic Indexing metrics, we propose a metric that highlights specific didactic aspects. This metric is based on the teacher's Teaching Styles namely, the Grasha teaching styles. Wikipedia pages are indexed with the Teaching Styles of the teachers that selected the given page; then, this value is compared with the Teaching Style of the teacher who is doing the search. The metric assigns higher rank to the Wikipedia pages that have a Teaching Style closer to the teachers'. We present an evaluation of the four retrieval metrics, showing that the teacher-styles based metric, Vs. the other three content-based metrics, allows for a better Wikipedia retrieval from a didactic point of view.

The public availability of datasets of teaching resources is an issue for the design, development and evaluation of Information Retrieval and Recommender Systems in Technology Enhanced Learning. Recently, the Dataset of Joint Educational Entities (DAJEE) has provided the community with a very exhaustive collection of resources coming from Massive Open Online Courses. This work proposes a representation of the DAJEE dataset according to Linked Data principles, interlinking the large amount of resources in DAJEE with the Web of Data. The transcript of the educational resources in DAJEE have been annotated through a Named Entity Recognition tool in order to create interlinks with the DBpedia entities. The DBpedia knowledge base provides additional information related to categories, that can be exploited to infer new knowledge and support reasoning processes.

The organization of the information in the knowledge economy has become a priority business process. Better organization leads to faster retrieval of relevant information. The process of searching and sequencing didactic materials for course building is an articulated and time-consuming process that requires considerable effort by the user. The goal of this research is to implement a platform for supporting the course building from Wikipedia articles. The selected materials will be automatically sequenced on the base of prerequisite relation, in order to provide a knowledge graph, editable by the teacher, with prerequisite relationships, representing the optimum learning path for the students.

With its 5,006,202 articles, 49 millions of registered people and on average 800 new articles per day, Wikipedia provides a knowledge base for teachers and instructional designers to build didactic materials. As a matter of fact, teachers often consult this encyclopaedia to arrange, integrate or enrich their courses. Moreover, with the exponential growth of the Internet, didactic materials are freely available and usable by teachers, instructional designers and students from Learning Objects Repositories such as Mertlot or Ariadne and others. On the other hand, designing and delivering a new course is a crucial task for teachers, who have to face two main problems: building or retrieving and sequencing learning materials. Retrieving or building learning materials requires a great effort and is time-consuming, while sequencing requires an accurate didactic project. In this paper we present a sequencing engine of learning materials, embedded in the Wiki Course Builder system, a system capable of retrieving and sequencing Wikipedia web pages, taking into account both the teacher model based on the Grasha teaching styles and on a social didactic approach. The main goal is to support teachers building on-the-fly courses, i.e., building courses quickly, with a few clicks of the mouse. An important feature of the system is represented by its ability to allow teachers to interact with the recommended learning path through a graph-based interface where they can directly modify the proposed learning path, adding or deleting Wikipedia pages. A first questionnaire has been submitted to a sample of teachers with encouraging results.

Selecting and sequencing a set of Learning Objects (LOs) to build a course may turn out to be quite a challenging task. In this paper we focus on such an aspect, related to the verification and respect of the relationships of pedagogical dependence existing between two LOs added to a course (meaning that if a given LO has another one as "pre-requisite", then any sequencing of the LOs in the course will need to have the latter LO taken by the learners before of the former). In our approach the sequencing of LOs in the course can still be managed by the instructor, basing on her/his taste and preferences, yet s/he can also be helped by a set of suggestions, related to the pre-requisite relationships existing among the LOs selected for the course. Such suggestions (such relationships, in effect) can be computed automatically and provide the instructor with significant help and guidance. We show a light-weight formalization of the LO, and how it can be "represented" by a set of WikiPedia Pages ("topics"); then we show how such set of topics, together with a set of relevant hypotheses we previously defined, can help establish the dependence relationship existing between two LOs. In this endeavor we exploit the classification in categories available for the WikiPedia topics, and obtain interesting results for our framework, in terms of precision and recall of the dependence relationships.

In the Technology Enhanced Learning (TEL) community, the problem of conducting reproducible evaluations of recommender systems is still open, due to the lack of exhaustive benchmarks. The few public datasets available in TEL have limitations, being mostly small and local.

In this paper we present a comprehensive framework supporting the tasks of defining, retrieving, and importing Learning Objects (LOs) for personalized courses. It is partially implemented in a Moodle-based personalization system, where the instructional designer is guided through: 1) a theoretical specification of the needed LOs; 2) a retrieval function of actual LOs, by automatically querying standard-compliant repositories; 3) an analysis of such items, to import those selected by him, also adding metadata relevant to the personalization system, at hand. This work overcomes some well known shortcomings of the Moodle system in supporting retrieval of learning material in a personalization context.

Curriculum sequencing is one of the most appealing challenges in Web-based learning environments: the success of a course mainly depends on the system capability to automatically adapt the learning material to the student's educational needs. Here we address the problem of how to compare and to test different curriculum sequencing algorithms in order to reason about them in a self-contained and homogeneous environment. We propose LS-LAB, a framework especially designed for comparing and testing different curriculum sequencing algorithms. LS-LAB has been designed to run different algorithms, each of them provided with its own student model representation: a super student model is able to incrementally include all of them. In this framework, the learning node has to be compliant to the IEEE LOM specifications, while, through a suitable GUI, one can insert new algorithms or run already available ones. We are carrying out the implementation by using a 3-tier Java application technology, in order to make this environment available on the Internet. Finally we show an application example.