Mr. Joel Rajnherc, an undergraduate student at the University of Aruba, successfully defended their Bachelor’s thesis entitled “Speech-to-text model for keyword spotting applications in the Papiamento language within a healthcare environment”, thereby achieving a momentous academic milestone within the Technology and Engineering specialization of the SISSTEM program.
During the defense on Friday, May 23rd, Joel demonstrated the results of a commendable research project that recorded audio from 280 volunteers at several locations, including the University of Aruba’s own campus and EPI. He then used the recordings to train a deep learning model to recognize spoken keywords in Papiamento. Joel used Convolutional Neural Networks, a type of neural network that works well on images, to detect patterns in image representations of audio data. The Bachelor’s thesis stood out as it is a first step at the University of Aruba, to guide our talented students in creating their datasets for future artificial intelligence applications that can help with accessibility, inclusivity, and offer new ways to interact with technology in our native language of Papiamento. “This is just the beginning of what’s to come. At the UA, we are dedicated to building the capacity on the island to address the challenges we face.”, said Dr. Sultan, dean of the Faculty of Arts and Science, and a co-evaluator of Joel’s thesis.
Joel narrowed down the area of research to healthcare, to explore how to integrate smart speakers in medical institutions such as ImSan in San Nicolas. Together with Full Stack Vision Foundation, he was also able to engineer the speech-to-text model to work in real-time on a Raspberry Pi 4 low-power computer. The thesis was supervised by MSc. Francis Laclé, and co-evaluated by Dr. Salys Sultan and Dr. Eric Mijts. Additionally, there were two external reviewers on the committee, MSc. Peter Scholing, who reviewed the digitalization aspect of Papiamento, and BSc. John Becker, who reviewed the research on the systems integration aspect of future devices in a hospital setting, such as ImSan. According to Becker, “By incorporating the native language of Papiamento, this technology will facilitate clearer communication with the patients”.
The successful defense of this thesis marks an essential step in Mr. Joel Rajnherc’s academic journey. It exemplifies the University of Aruba’s commitment to helping students contribute to positive and sustainable innovations in the rapidly evolving field of artificial intelligence.
