AI-based platform for enhancing and personalizing e-learning

Researchers on the Autonomous University of Madrid lately created an revolutionary AI-powered platform that may improve distance studying. This permits educators to securely monitor college students and be sure that college students are collaborating in necessary on-line lessons and exams.

The first prototype of the platform, referred to as Demo-edBB, can be unveiled on the AAAI-23 Conference on Artificial Intelligence in Washington in February 2022. arXiv preprint server.

“Our analysis group BiDA-Lab Dr. Roberto Daza Garcia of Universidad Autonoma de Madrid informed TechXplore:

“Over the past few years, virtual education has grown significantly, becoming a major platform for one of the most important educational institutions and creating new and valuable opportunities for learning. We are working on new technologies in learning and eventually leading the way to developing a platform that combines biometrics and behavioral analysis tools.”

EdBB, a platform created by the BiDA-Lab staff, was particularly designed to enhance the web pupil evaluation course of whereas enhancing safety. The platform contains a number of biometric instruments and pre-trained algorithms that acknowledge customers based mostly on their habits (reminiscent of keyboard and “keystroke” utilization patterns) or physiological knowledge (reminiscent of facial recognition instruments). based mostly on expertise. Detect particular behaviors (consideration, stress, and so forth.). So far, the researcher has developed a demo of his model of the platform referred to as edBB-demo, however is presently engaged on an built-in model.

“Our platform captures numerous sensors (webcam, keyboard, audio, metadata, and so forth.) from the common pupil’s laptop and applies numerous applied sciences to the applying. real timeestablish customers, suspicious occasions, habits presumptions, and so forth., and then define them in a report for lecturers,” defined Daza Garcia.

“We can capture all student sensors in a secure and transparent way, while allowing students to use other online education platforms. We are combining some of the advances.”

The platform created by this staff of researchers depends on a multimodal studying framework, a mannequin that may analyze various kinds of knowledge reminiscent of photographs, movies, audio alerts and metadata. A demo model of the platform was educated on a database of studying and testing classes. Each session he lasted over 20 minutes and was attended by 60 completely different college students.

“One of the biggest concerns for educational institutions is how to prove that remote students are actually participating in online assessments,” stated Daza Garcia. “edBB-Platform’s biometrics and habits detection expertise ensures better safety on this necessary problem, whereas detecting pupil habits to enhance the training course of and estimate pupil consideration and stress ranges. We imagine that these new applied sciences can be elementary to offering a extra individualized schooling for every pupil sooner or later.”

The demo model of edBB has 4 necessary options. This means authenticating customers with a excessive degree of accuracy, recognizing human actions in movies, utilizing webcam footage to estimate a pupil’s coronary heart fee, and analyzing facial expressions to estimate a pupil’s consideration. can do. The dataset used to coach the framework has lately been made out there on-line, so it may be used to coach different machine studying fashions.

The platform created by this staff of researchers might quickly assist advance distance learning, permitting educators to reliably and securely confirm the identification of eLearners. Additionally, it could facilitate personalization of on-line studying by figuring out points which may be hindering pupil studying, reminiscent of decreased consideration span or excessive stress ranges.

“We want to continue improving edBB because we believe this is a large area with a promising future that faces many challenges.platform“We would like to continue developing the line of research we are currently working on. We also want to develop a novel cognitive load that uses multimodal face analysis and novel multimodal architectures to identify keyboard or mouse dynamics in students. We also want to develop an estimation system, and we would like to expand our research areas to include visual attention estimation, eye-tracking, and response prediction.”

© 2022 Science X Network