Researchers at Universidad Autónoma de Madrid have developed an innovative, AI-powered tool that might improve remote learning by enabling instructors to safely monitor students and confirm that they are attending mandatory online lectures or tests. An early version of this platform, dubbed Demo-edBB, will be presented at the AAAI-23 Conference on Artificial Intelligence in Washington in February 2022, and a version of the work is accessible on the arXiv preprint service.
“Our research group, the BiDA-Lab at Universidad Autónoma de Madrid, has extensive expertise with biometric signals and systems, behaviour analysis, and AI applications, with over 300 hundred published publications in the previous two decades,” said Roberto Daza Garcia, one of the study’s authors.
“Virtual education has risen dramatically in recent years, becoming the primary basis of one of the most prominent educational institutions and producing new excellent learning possibilities. As a result, our group has lately been working on new e-learning technologies, eventually leading to the creation of a platform that integrates biometric and behaviour analytic capabilities.”
The BiDA-Lab team’s platform, EdBB, was particularly built to enhance online student assessment procedures while simultaneously increasing their security. The platform is built on a variety of technologies, including biometric identification tools that recognise users based on their behaviour (e.g., patterns in keyboard use or “keystrokes”) or physiological data (e.g., facial recognition tools), as well as algorithms that have been trained to detect specific behaviours (e.g., attention, stress, etc.). The researchers have so far created a trial version of their platform, nicknamed edBB-demo, but they are currently working on the full version.
“Our platform takes various sensors from the ordinary student’s computer (webcam, keyboard, audio, metadata, and so on) and uses multiple technologies in real-time to identify users, suspicious occurrences, behaviour estimates, and so on,” Daza Garcia added.
“It can securely and transparently record all students’ sensors while letting students to utilise any other online education platform. edBB-Demo brings together some of the most significant breakthroughs in remote biometric and behavioural knowledge over the previous decade.”
This research team’s platform is based on a multi-modal learning framework, a model that can assess many forms of data such as photos, videos, audio signals, and metadata. The platform’s trial version was trained on a database of learning and assessment sessions spanning more than 20 minutes and involving 60 distinct students.
“One of the most pressing problems for educational institutions is proving that distant students are really participating in an online examination,” Daza Garcia stated. “The edBB-biometric Platform’s and behavioural detection technologies may increase security while also identifying a student’s behaviour, which might enhance the learning process and possibly open the path for new technologies to assess students’ attention or stress levels. We are certain that these new technologies will be critical in the future in order to provide more individualised education for each student.”
The demo version of edBB can authenticate users with high accuracy, distinguish human behaviours in films, estimate a student’s pulse rate using webcam footage, and assess a student’s attentiveness by studying their facial expressions. The dataset used to train the framework was recently made publicly accessible online, and so might be used to train other machine learning models.
This team of researchers’ technology might soon assist to revolutionise remote learning by enabling educators to reliably and securely authenticate the identification of e-learners. Furthermore, it may aid in the customization of online learning by recognising potential difficulties impeding a student’s learning, such as low attention or excessive stress levels.
“We feel this is a broad sector with a great future but many hurdles ahead, thus we now want to continue enhancing the edBB-platform,” Daza Garcia stated. “We intend to continue developing the present research lines as well as new cognitive load assessment systems that employ multimodal face analysis and new multimodal architectures to recognise the student’s keyboard or mouse movements. Furthermore, we wish to broaden our research areas to include visual attention estimates, gaze tracking, response prediction, and so on.”
You might also be interested in reading, What Is The Reason Behind ChatGPT’s Unbelievable Number Of Users? Costs are Eye-watering