Conversational agents are increasingly used in education, yet most educational chatbots remain content-centric and rely on learner models that represent knowledge states while leaving the learning process itself implicit. This paper introduces a process-oriented learner model that embeds principles from Agile methodology directly into the internal state of an AI-driven educational chatbot to support students' independent learning. In this model, learning objectives are represented as epics, study cycles as competence-driven sprints, and micro-level goals as user stories governed by explicit definitions of done. We implement the approach in a multi-agent Telegram chatbot used in an introductory university programming course. An exploratory feasibility study with 16 university students and recent graduates (N=16) provides preliminary evidence of the approach's pedagogical promise. Participants reported that the Agile-structured interaction improved goal clarity, step-by-step progressio n, and reflective engagement, though technical limitations in response latency and explanation depth were identified. This paper shows the feasibility of embedding explicit process models within conversational AI for education and motivates further investigation with larger samples and controlled evaluations.

Artificial intelligence promises to revolutionize our life, bringing significant advances in any fields: health, education, work and leisure time. This paper analyzes the last 5-year literature concerning the use of AI for supporting people with disabilities in education. The aim is to investigate the current state of art of accessible applications in the STEM (Science, Technology, Engineering, Mathematics) field and understand if contents and tools are accessible for all, regardless of personal need and abilities. Personalization and adaptation emerge as fundamental factors when designing for people with disabilities. Privacy and ethics aspects often neglected are very relevant. The analysis suggests that the STEM field still suffers from accessibility gaps, and current tools need to evolve and be increased to be exploited by different disabilities and ensure the same opportunities for every student, engaging, motivating, and empowering them..

Virtual Learning Environments (VLE) are present in all educational modalities and levels, but their adoption in primary education remains limited. This article evaluates the acceptance of MundoCRA, a VLE that enriches traditional classrooms through a virtual world where students perform educational tasks in an engaging way. Gamified activities with rewards, points, and certificates promote motivation through quizzes, challenges, and collaborative projects. MundoCRA also includes a management module that allows teachers to create and monitor tasks easily. The evaluation, based on an extended Technology Acceptance Model (TAM), considers constructs such as perceived enjoyment, satisfaction, and gender differences. A survey of 20 teachers and 30 students from the Rural Grouped School (CRA) Sierra de Alcaraz (Spain) revealed that MundoCRA increased motivation, satisfaction, and enjoyment. Furthermore, structural equation modeling showed that perceived enjoyment and ease of use significantly influenced positive attitudes toward the platform, explaining over 70% of the variance in user attitudes.

Alternative descriptions of digital images have always been an accessibility issue for screen reader users. Over time, numerous guidelines have been proposed in the literature, but the problem still exists. Recently, artificial intelligence (AI) has been introduced in digital applications to support visually impaired people in getting information about the world around them. In this way, such applications become a digital assistant for people with visual impairments. Increasingly, generative AI is being exploited to create accessible content for visually impaired people. In the education field, image description can play a crucial role in understanding even scientific content. For this reason, alternative descriptions should be accurate and educational-oriented. In this work, we investigate whether existing AI-based tools on the market are mature for describing images related to scientific content. Five AI-based tools were used to test the generated descriptions of four STEM images chosen for this preliminary study. Results indicate that answers are prompt and context dependent, and this technology can certainly support blind people in everyday tasks; but for STEM educational content more effort is required for delivering accessible and effective descriptions, supporting students in satisfying and accurate image exploration.

The paper presents a comparative mixed methods study of the accessibility and usability for disabled people of four video conferencing tools, Zoom, MS Teams, Google Meet and Skype. Useable responses were obtained from 81 disabled people with diverse characteristics, mainly in the UK, though some groups had low representation. None of the tools was considered fully accessible and useable. Zoom was both the most commonly used and the most frequently preferred (56.1%) tool, with MS Teams second in usage and a trailing second in preferences (15.9%). It was considered to have better captioning, but otherwise to generally be a poor second to Zoom. Skype was the most commonly used before lockdown, but was considered dated and its limited use was mainly social, whereas the other tools were also used in work and education. The results were used to draw up separate lists of recommendations for developers and meeting organisers and hosts, as the study also identified actions for organisers and hosts to improve meeting accessibility. Developer recommendations include several easy to set customisation and user friendly interface features, involving disabled people and specific accessibility features, including compatibility with assistive technology, keyboard shortcuts for all functions and automatically-on high quality captions.

Abstract Breast cancer (BC) is known as the most prevalent form of cancer among women. Recent research has demonstrated the potential of Machine Learning (ML) techniques in predicting the five-year BC risk using personal health data. Support Vector Machine (SVM), Random Forest, K-NN (K-Nearest Neighbour), Naive Bayes, Neural Network, Decision Tree (DT), Logistic Regression (LR), Discriminant Analysis, and their variants are commonly employed in ML for BC analysis. This study investigates the factors influencing the performance of ML techniques in the domain of BC prevention, with a focus on dataset size and feature selection. The study's goal is to examine the effect of dataset cardinality, feature selection, and model selection on analytical performance in terms of Accuracy and Area Under the Curve (AUC). To this aim, 3917 papers were automatically selected from Scopus and PubMed, considering all publications from the previous 5 years, and, after inclusion and exclusion criteria, 54 articles were selected for the analysis. Our findings highlight how a good cardinality of the dataset and effective feature selection have a higher impact on the model's performance than the selected model, as corroborated by one of the studies, which gets extremely good results with all of the models employed.

Abstract Playing games is an important way to promote the integration, inclusion, and socialization of participants. This is especially the case of persons with disabilities, such as visually impaired people. Unfortunately, very few games are accessible to such persons. Hopefully, in many digital games, this accessibility can be enabled in principle by assistive technologies, such as screen readers. The aim of this work consists in reviewing the recent literature on game accessibility for people with visual impairment and discussing benefits, limitations, and possible improvements of currently available accessibility solutions. After providing a definition of visual impairment and describing its relationship with gaming, the work reviews general techniques for designing more accessible games. Subsequently, it focuses on specific techniques based on replacing visual stimuli with auditory stimuli (e.g., sonification and sound-source simulation), also presenting some recently proposed sonification-mapping strategies. Then, the application of machine-learning techniques to the development of accessible interfaces for online versions of board games is illustrated by a recent case study. Finally, a discussion and some conclusions are provided, with a particular focus on policy implications of improvements in game accessibility for visually impaired people.

Nowadays mobile devices are essential tools for visiting cultural heritage sites. Thus, it is very important to provide an inclusive cultural mobile experience for anyone. In this study, we investigate how to create accessible apps to enhance the experience of the visually impaired people in outdoor cultural itineraries. In such a context, the integration of specific features for improving accessibility for blind people may require high skills. This study investigates how it is possible to exploit a simple-to-use developing environment, Thunkable framework, which does not require specific technical competencies, to easily develop an accessible app.

Independently selecting food items while shopping, or storing and cooking food items correctly can be a very difficult task for people with special needs. Product labels on food packaging contain an ever-increasing amount of information, which can also be in a variety of languages. The amount of information and also the features of the text can make it difficult or impossible to read, in particular for those with visual impairments or the elderly. Several tools or applications are available on the market or have been proposed to support this type of activity (e.g. barcode or QR code reading), but they are limited and may require the user to have specific digital skills. Moreover, repeatedly using an application to read the label contents can require numerous steps on a touch-screen, and consequently be time-consuming. In this work, a portable tool is proposed to support people in reading the contents of labels and acquiring additional information, while they are using the item at home or shopping at the supermarket. The aim of our study is to propose a simple portable assistive technology tool which 1) can be used by anyone, regardless of their digital personal skills 2) does not require a smartphone or complex device, 3) is a low-cost solution for the user.

EPUB (Electronic Publication) is a format primarily used for digital books. EPUB 3 is based on the Open Web Platform, HTML, CSS and JavaScript. In short, EPUB content can be considered as a web interface. Thus, existing accessibility guidelines and techniques can be applied in order to obtain an accessible eBook. In our study, we investigate if popular editing tools are sufficiently mature to support the authors in designing interactive and multimedia EPUB 3 fully accessible via assistive technology like screen readers. To this purpose, in this work, the Adobe InDesign tool has been used to prepare an interactive EPUB 3 prototype as a case study. Our aim was to analyze whether the InDesign tool was able to support authors in getting a fully accessible interactive EPUB 3. The results revealed that accessibility is not fully guaranteed directly by the tool: some additional steps are required by the authors or professional operators.

Home Automation Systems (HAS) -also referred to as smart homesexploit multiple components such as sensors, RFID readers, wireless devices, and remote control systems to enable easy interaction with smart appliances and devices, and to automate performing sequences of tasks to make human-device interaction simpler and life more comfortable.For people with vision impairment, especially those who are unable to see at all, smart homes can be a powerful tool for enhancing personal autonomy, provided that the system offers suitable device integration and accessible interfaces with a simple interaction via keyboard, assistive technology and other modalities such as voice and gestures.This paper investigates the accessibility of web interfaces when interacting with HAS components via screen reader assistive technology, in order to propose potential suggestions to developers.

Last-decade great advances in technology have contributed to make home smarter and more comfortable, especially for people with disabilities. A lot of low cost solutions are available on the market, which can be controlled remotely by a Home Automation System (HAS). Unfortunately, the user interfaces are usually designed to be visually oriented which can exclude some user categories, like those who are blind. This paper focuses on the design of usable Web user interfaces for Home Automation Systems, with a special attention to the functions as well as the interface arrangement in order to enhance the interaction via screen reader. The proposed indications could inspire other designers to make the user experience more satisfying and effective for people who interact via screen reader.

In this report, we share our experience and observations on the challenges blind people face with text entry on touch-based mobile phones, particularly from the perspective of one of the authors, who is blind. To better understand these issues we developed and tested Multimodal Text Input Touchscreen Keypad (MTITK), an audio-tactile text entry prototype based on multitap, which relies on a telephone keypad layout organized into five key groups with distinct audio-tactile feedback. Users explore the screen to identify the current selected key, tap to enter text, and gesture to edit it, while receiving the corresponding voice, audio, and tactile feedback; no additional equipment is necessary in our software-only approach. We implemented a prototype on Android and tested its usability with visually impaired participants; they welcomed its multimodality and the familiar layout, but also expressed the need to increase vibration pattern differentiation and refine the character selection mechanism.

This tutorial discusses accessibility and usability issues for disabled students. It briefly reports on a study of disabled students and teachers in the UK, Italy and Poland and presents guidelines and recommendations.

Social interaction is important for everyone, and vital for the differently-abled and elderly. Social network applications are causing us to change our ways of communicating and working. We can now make friends all over the world, share ideas and knowledge, search for others' opinions, and more. Facebook and Twitter are examples of these applications. However, these opportunities are usually meant for people who can walk, see, and talk without difficulty. In this paper we show how blind persons, using screen readers and voice synthesizers to access the Web, can interact with social network applications. Our case study focuses on Facebook, analyzing basic features, explaining electronic barriers and suggesting solutions.

Audio podcasting is increasingly present in the educational field and is especially appreciated as an ubiquitous/pervasive tool ("anywhere, anytime, at any pace") for acquiring or expanding knowledge. We designed and implemented a Web-based Text To Speech (TTS) system for automatic generation of a set of structured audio podcasts from a single text document. The system receives a document in input (doc, rtf, or txt), and in output provides a set of audio files that reflect the document's internal structure (one mp3 file for each document section), ready to be downloaded on portable mp3 players. Structured audio files are useful for everyone but are especially appreciated by blind users, who must explore content audially. Fully accessible for the blind, our system offers WAI-ARIA-based Web interfaces for easy navigation and interaction via screen reader and voice synthesizer, and produces a set of accessible audio files for Rockbox mp3 players (mp3 and talk files), allowing blind users to also listen to naturally spoken file names (instead of their spelled-out strings).

Link annotation with the accessibility level of the target Web page is an adaptive navigation support technique aimed at increasing blind users' orientation in Web sites. In this work, the accessibility level of a page is measured by exploiting data from evaluation reports produced by two automatic assessment tools. These tools support evaluation of accessibility and usability guideline-sets. As a result, links are annotated with a score that indicates the conformance of the target Web page to blind user accessibility and usability guidelines. A user test with 16 users was conducted in order to observe the strategies they followed when links were annotated with these scores. With annotated links, the navigation paradigm changed from sequential to browsing randomly through the subset of those links with high scores. Even if there was not a general agreement on the correspondence between scores and user perception of accessibility, users found annotations helpful when browsing through links related to a given topic.

In this paper, we present a solution for supporting vibrotactile feedback in mobile museum guides for blind users. To this end, we have designed and implemented a hardware/software module, which can be easily plugged into current PDAs to assist blind users in orientation. The solution, which comprises a two-channels haptic module as well as vocal support, has been exploited for moving through tagged objects. We also report on a user evaluation carried out with a number of blind users.

Available accessibility guidelines do not necessarily guarantee usable Web sites, particularly when specific groups of users with special needs are considered. We have identified 15 Web design criteria aiming to provide integrated support of accessibility and usability for vision-impaired users. In this article, we present the results of a study investigating whether the application of such guidelines for vision-impaired users can actually improve their task performance when accessing Web applications. We report on two user tests, both involving vision-impaired users, that aim to provide empirical validation of the design criteria. During each test, users had to access and navigate two versions of a Web site, one version supporting the selected design criteria and one obtained with traditional techniques. Our results indicate that the 15 design criteria improved Web site usability both quantitatively and qualitatively by reducing the navigation time needed to perform the assigned tasks and by making the Web sites easier to navigate for blind and low-vision users.

In this study we present a SW prototype developed within the framework of a research project aiming at improving the usability of search engines for blind users who interact via screen reader and voice synthesizer. Following the eight specific guidelines we proposed for simplifying interaction with search engines using assistive technology, we redesigned Google user interfaces (i.e. simple search and result pages) by using XSL Transformations, Google APIs and PERL technologies. A remote test with 12 totally blind users was carried out in order to evaluate the proposed prototype. Collected results highlight ways in which Google interfaces could be modified in order to improve usability for the blind. In our demo we will show how interaction with the modified Google UIs is simplified and how the time for reaching the most important elements (i.e. first query result, next result page, etc.) is shortened in comparison to interaction with the original Google UIs. The demo uses the JAWS screen reader for announcing the UI contents.

We present the results of a research project focus on improving the usability of web search tools for blind users who interact via screen reader and voice synthesizer. In the first stage of our study, we proposed eight specific guidelines for simplifying this interaction with search engines. Next, we evaluated these criteria by applying them to Google UIs, re-implementing the simple search and the result page. Finally, we prepared the environment for a remote test with 12 totally blind users. The results highlight how Google interfaces could be improved in order to simplify interaction for the blind.

The increasing need to check Web site accessibility has stimulated interest in tools to aid the various activities involved. While some tools for this purpose already exist, we believe that there is a demand for making their support more flexible. In particular, there is often a need for validation of multiple sets of guidelines, repairing Web pages and providing better reports for the evaluators. In this paper, we discuss such issues and how we have addressed them in the design of MAGENTA, our new tool for supporting inspection-based evaluation of accessibility and usability guidelines.