Interest in Artificial Intelligence captivated Dorin Doncenco: 'I genuinely get to work on the latest technology'
In the autumn of 2020, Dorin Doncenco, an ICT student at Turku UAS, studied in a course on artificial intelligence. On the course, he became acquainted with deep learning, and towards the end of the course he was able to study deep learning in radiology in a practice-oriented manner.
At the end of the course, the teacher, Mojtaba Jafaritadi, asked Dorin if he was interested in deepening his Artificial Intelligence skills.
“He offered me the opportunity to join the PRIVASA project. I am currently working on the project as a student assistant. In addition, I do my thesis on the topic” Dorin says.
At the same time Dorin is also completing his internship in the PRIVASA project.
Working on the latest technologies
Turku UAS is involved in the PRIVASA — Privacy Preserving AI for Synthetic and Anonymous Health Data project to accelerate product development of companies operating in international markets by producing anonymous, individual-level and synthetic health data.
The project is a Co-Innovation project funded by Business Finland and coordinated by the University of Turku. The project is implemented by the Health Technology research group at Turku University of Applied Sciences.
In the project, Turku UAS is developing artificial intelligence methods which guarantee data protection for processing sensitive health and imaging data. Machine learning algorithms built at Turku UAS produce synthetic data based on various actual medical records.
Dorin's job description includes a lot of reading and learning about various studies. What he learns he applies in health technology. Upon discovering valuable information, he reports it to the other members of his team and those working on PRIVASA.
“After starting out as a student assistant, I learned a lot about GAN, which is generative adversarial networks. I read a lot of articles in the field, after which I apply what I’ve read in practice”, Dorin says about his work.
A generative adversarial network (GAN) consists of two artificial intelligence systems trained compete with each other and create more realistic data than a system that works alone.
The system performing the work phase continuously learns from the feedback it receives. The same will continue for as long as the outcome is satisfactory. GANs don't just memorize past content; they can create entirely new content as well.
“GAN and its methods are reasonably new and the methods of synthetic data are a really hot research topic at the moment”, says Elina Kontio, the head of the Health Technology research group.
Dorin says he often stumbles upon new things about which there isn't yet much information. His job description is to solve oncoming problems — with team support and sometimes even alone. GAN was developed in 2014, and in his work, Dorin will read articles on the subject that have been published over the next few years.
“I genuinely get to work on the latest technologies. Before this job, I imagined that the phrase in question is just an advertising gimmick for companies, but this time it is true”, he rejoices.
Graduation in the horizon
Dorin hails from Moldova but has been studying with a double degree since 2017, aiming for a bachelor's degree in computer science at Western University Timisoara, Romania. In addition, he has supplemented his education with ICT studies at Turku UAS.
He is due to graduate this coming spring, as the only things left to complete are the practical training and thesis.
Interest in AI has captivated Dorin, and he hopes to become employed in the field in the future. Also, in his postgraduate studies, he plans to include AI.
“Artificial intelligence is very new and complex. I would like to complete a Master’s degree on the subject in order to gain a better understanding of the field so that I can become more involved in project activities”, he says.
Besides she is a leader of the research group, Kontio also is the instructor of Dorin’s internship. According to her, Dorin has shown activity and developing in learning methods of AI.
“It is very great to work with an enthusiastic student and to get experts advanced as part of a challenging AI project”, Kontio sums up.
In addition to goal-oriented studying, his life also accommodates a bunch of hobbies.
“My hobbies change regularly. In between, I read books, at times I play chess at the amateur level. I'm interested in science in my free time, but I also read a lot of articles on things like religion and psychology. To counter my hectic, work- and study-filled everyday life, I've also started meditating”, he says.