It was recently announced that the Research and Development Institute for Artificial Intelligence of Serbia, which will deal with research in the field of artificial intelligence and machine learning, will start operating in June in the Science and Technology Park in Novi Sad. His chief researcher, Branislav Kisacanin, sees it as “a factory of artificial intelligence leaders, which will conquer a part of the market for Serbia that may soon be worth 500 billion dollars a year.”
Startit journalists talked to experts about how the Institute is conceived, who will run it, who it will be intended for and how it will fight not to become a sluggish state institution.
“It is a state-supported startup, which for now has its own skeleton and vision,” explains the acting president, Dubravko Culibrk, professor and director of the Department of Industrial Engineering and Management at the Faculty of Technical Sciences in Novi Sad. He sees the Institute as a place that will provide domestic researchers with the best framework for development.
He adds that machine learning is now taught in numerous study programs, but that none is dedicated to just that, and that the Institute should be a focal point, where everyone will be focused only on that area, because that is the way to develop it. He explains how good it is that this institution will not belong to any university and their plan is to have their departments in several cities.
“We hope that we will not become a sluggish state institution. The institute must constantly show innovation. Our goal is to attract investments with our work and become self-sustainable,” assures Culibrk.
It is easier to get staff through the Institute than through universities
For the economy, the Institute should be a place where staff is created, says Filip Panjevic, president of the Board of Directors and a representative of domestic business, as a co-founder of ydrive.ai.
“We now do not have enough staff who have passed through educational institutions and are ready to work, but they are also educated through work in the economy. It would be ideal for universities to do that, but the influence of machine learning in curricula is a long-term process. I think that in the short term, it is easier to speed up that educational process in a smaller institution, than to reform entire universities,” says Panjevic.
Hard to get to the first researchers?
However, investment is not the main problem. At the moment, it seems that the most difficult aspect is to find adequate staff, ie the first researchers. Much depends on their quality, because the first employees will build the culture of this institution. Chief researcher Branislav Kisacanin, who received his doctorate from the University of Illinois and has been working in the industry for years, is mainly in charge of finding these people, precisely on the development of computer vision and machine learning.
He now lives and works in America, he stays in Novi Sad longer and longer, and his wish is that when the Institute develops, it will become his main focus. As he says, he wants to return to Serbia, as much as possible, what it has invested in it, by mediating that top knowledge is brought to our country. Similar reasons encouraged his colleagues, originally from Serbia, who now work around the world, to join the mentoring program, as they designed.
“For the first two years, the Institute will be like postdoctoral studies, where we will help our researchers raise the level of their research to a world level. In the beginning, we will work on such projects that are focused on the development of science, and soon those that are needed by the economy. To do this, we have created a mentoring program with our people from around the world, which will help us in all aspects, from choosing research areas, through publishing papers in relevant publications, to designing new projects and winning funding from various funds, describes Kisacanin.
We understand that researchers will go into the economy, alumni are important
Such a program is one of the things that attracts future candidates the most to work at the Institute, in addition to the fact that they will be able to do science. At the beginning, there was a fear that there would be enough quality young researchers, who are ready to be the first generation. Kisacanin says that he has already received a dozen quality biographies, so that what initially seemed like a challenge, he now sees as a task.
The Institute will be able to provide them with a decent income, but it cannot compete with large companies, everyone at the Institute is aware of that. As Kisacanin says, one of the main trump cards with which the Institute attracts talents is that they will be able to learn a lot in it and after two or three years they want to establish their own companies or move to universities or existing companies.
“It is actually inevitable at a time when the demand for such personnel is huge, so the Institute has incorporated that principle into its DNA. Namely, it will function as a flow boiler for talents. Instead of regretting when employees leave us one day, we will even help them establish startups and patent discoveries and inventions. We want to stay in close collegial relations with all of them and for them to become the next generation of mentors to those who succeed them at the Institute. On the other hand, some will really like the atmosphere at the Institute and will stay with us for a long time. Of course, that is also good, because larger projects will need such continuity.
When it comes to areas, they will try to do everything that is current, from computer vision, language modeling, application in health care, and they will also have a group that will deal with the mathematical foundations of machine learning. Mentors are also there to help in which directions and areas the research will go, in order to avoid dead ends.
“For now, we have no plans to conduct research in the field of ethics in the application of these technologies. This does not mean that not every researcher should think carefully about it, because the consequences can be serious. We all watched Terminator. When the Institute is sufficiently developed, we will probably open several positions in the field of ethics. In that sense, the Institute will deal with areas that are pure civilizational gain, for example, the further development of machine learning for the design of new drugs and materials,” says Kisacanin.
Everything they are working on will be shared with the public, which is a feature of AI researchers and why, according to Kisacanin, this area is developing exponentially. The community itself is very open, and even companies publish a large part of their results and data.
We have the task to prove that the Institute is not a trick
Everything that our interlocutors talked about, those biographies of people who have been involved voluntarily for now, seems correct for now. However, there is skepticism because it is a state institution, so as Filip Panjevic says: “I think it is a big task for us to prove that the Institute is an institution that makes sense and deals with top science, and is not a trick,” BiF reports.
