A Platform to Support Personalized Training of People with Disabilities
Carlos Cardonha, Andrea Britto Mattos, Rodrigo Laiola Guimarães · 2016 · Proceedings of the 13th International Web for All Conference (W4A) · doi:10.1145/2899475.2899493
Summary
This paper from IBM Research Brazil presents a two-component platform for creating and delivering personalized vocational training courses to people with multiple types of disabilities. The system consists of a web-based course-authoring portal for instructors and a tablet-based mobile application for students. The platform addresses a significant gap at the intersection of assistive technology and vocational training, where the authors found surprisingly little prior research. The key design insight is that personalization for learners with disabilities requires not just selecting appropriate content, but providing alternative representations of the same content — for example, presenting a step as plain text for a learner with hearing impairment and as audio for someone with visual impairment. Courses are structured as sequences of presentation steps (combining text, video, image, and audio via templates) and evaluation steps (multiple-choice quizzes). Critically, courses use a non-linear directed graph structure where quiz answers can branch to different paths, allowing instructors to route students who answer incorrectly to alternative explanations before returning them to the main sequence. A tag-based system (visual impairment, hearing impairment, cognitive impairment, no impairment) lets instructors assign accessibility profiles to students and create tag-specific compositions for each step, with the mobile app automatically delivering the appropriate version.
Key findings
A preliminary pilot with two experienced instructors (6+ years each) from APAE DE SAO PAULO, a Brazilian institution providing supported employment and vocational training for people with intellectual disabilities, showed favorable reception. Both instructors found the web portal intuitive and completed all authoring tasks successfully. They highlighted three features as particularly valuable: (1) The non-linear branching structure, which lets them determine whether a student missed a question due to an intellectual limitation or due to the content presentation mode — a distinction critical for effective teaching of people with ID. (2) The tag-based personalization, with one instructor noting she could use it to deliver larger images for students with Down syndrome who have visual limitations alongside other content formats for students with different types of ID in the same classroom. (3) The progress logging, which records time spent on each step and quiz answers in real-time — one instructor noted this could reveal whether difficulties stem from laziness or anxiety rather than cognitive limitations, information currently only available through direct observation. The platform also supports both self-paced and instructor-mediated execution modes, the latter addressing observed challenges with student attention during MOOC-based learning.
Relevance
This work highlights the largely underexplored intersection of vocational training and accessibility technology. While assistive technologies for people with disabilities are well-researched, their application to employment training — which directly affects economic independence and inclusion — receives far less attention. The platform's approach to personalization through alternative content representations rather than simplified content is an important distinction: it maintains the same learning objectives while adapting presentation to individual needs. For accessibility practitioners and organizations, the tag-based system demonstrates a practical, instructor-driven approach to personalization that acknowledges the current lack of data needed for fully automated adaptation. The study's context in Brazil's supported employment ecosystem also provides a valuable non-Western perspective on disability services and vocational inclusion. The small pilot scale (two instructors, no student evaluation) limits generalizability, but the architectural approach to multi-disability content delivery is broadly applicable.
Tags: intellectual disability · educational technology · personalization · vocational training · mobile accessibility · inclusive design