Dubravko Ćulibrk Ph.D., Acting Director Of The Institute For Artificial Intelligence R&D Of Serbia

The Artificial Intelligence Development Strategy and action plan have proven to represent a realistically envisaged route for Serbia to enter the field of AI applications. Establishing the Institute was the most ambitious element, so its very existence provides proof of how realistic the strategy is

Artificial intelligence represents technology that will bring future change to many aspects of life and business. If we succeed in understanding, accepting and applying this technology in the right way, we have a chance to improve the standard of living for every citizen of Serbia, but also of the region and even the world. This is the notion that guides the work of our interlocutor, Dubravko Ćulibrk Ph.D., acting director of the Institute for Artificial Intelligence R&D of Serbia, who has been at the helm of this institute since it was founded. We caught up with him to discuss the Institute’s two years of work to date and its further development.

It seems that there is growing awareness here in Serbia of the need for the domestic economy to be based on knowledge and innovation. How much do you recognise these tendencies in the actual milieu of the economy and in which areas?

– My entire career has been focused on information and communications technology, and in that part of the economy there is a notable trend of development from services towards business oriented to the development of vendor products and – particularly over the last few years – towards the research, creation and commercialisation of intellectual property.

Here I’m referring to the broader field of computing-based operations and electrical engineering, which includes areas like robotics, mechatronics and parts of the creative industries. It seems to me that we can also expect these kinds of development tendencies in the field of biotechnology soon.

Where do you see the AI Institute’s place in this process?

– We are a research & development institute and one of the basic elements of our vision is to become a mint that forges leaders in the domain of artificial intelligence. In practical terms, this means that we want to produce PhDs who possess research experience and knowledge at a level considered relevant on the global labour market, but also to educate them on how to protect the results of their research, and subsequently how to commercialise those results, through licensing or the creation of start-up companies. In this regard, our institute is – to the best of my knowledge – the only institution of this kind in the region to have an intellectual property protection policy and that maximally favours the people who created that intellectual property, while of course respecting legal limitations.

An important aspect of the mission of our institute is to create an environment of worldclass excellence for researchers dealing with artificial intelligence in our country and to support the development of their personal excellence… We have endeavoured to set a regional standard in this regard from the very outset, and it seems to me that we’ve succeeded

We have implemented this through clear guidelines that regulate the protection of intellectual property at the institute, as well as the conditions under which the Institute supports the forming of new companies on the basis of that property. Moreover, within the scope of the Institute we also have the operational capacity to write patents, which isn’t the norm among our scientific research institutions. In this sense, we plan to lead by example in the process of developing a knowledge-based economy and we are working actively to build an ecosystem of deep tech entrepreneurship in the domain of artificial intelligence and its applications.

To what extent have the Artificial Intelligence Development Strategy and associated action plan proven to represent a realistically envisaged route for Serbia to enter the field of AI applications?

– The Strategy relies on several pillars. I think the strategy and action plan have proven extremely realistic when it comes to education and research. In the case of the latter, I would say that the founding of the Institute was the most ambitious element, thus its very existence provides proof of how realistic the strategy is. It is yet to be unequivocally demonstrated how realistic it is in the field of AI applications in the public sector and the development of an AI-based economy, but my personal opinion is that the effort exerted in this regard should be justified by the end of the period covered by the strategy and action plan.

Two years after the establishment of the Institute, do you think it’s on track to becoming a global centre of excellence?

– I think it is. I never thought that would be easy or that success was guaranteed, but I would say that now, after two years, we already represent a regional centre of excellence. That’s a good result for such a short period. We have established an organisation capable of implementing our vision and have a clear strategy for reaching our goal.

Regardless of the fact that we still have a long way to go, when I look at our advisory board, I’m truly amazed at what a global “centre” of excellence it represents. These kinds of people wouldn’t waste their precious time dealing with something that has no prospects of becoming relevant globally.

To what extent is the thematic division of the work of your research groups determined by global trends in AI applications and to what extent does it depend on the potential needs of the domestic economy? Within the scope of the domestic economy, who are your main interlocutors when it comes to implementing AI-based solutions?

– Our groups are a mix of expertise that we were able to find in our country and our economy’s strategic development priorities. They are thereby fully aligned with global trends, though, of course, we lack the capacity to engage in research related to every global trend. We ended up establishing five groups. Two are focused on core technologies that deal with human-machine interaction and the field of computer vision, while three deal with applications that are attractive globally and locally: in medicine and the life sciences, smart manufacturing and environmental protection and responses to climate change. Many of our partners are the local representative offices of global companies that invest significant resources in the development of new technologies, amongst which is Japanese innovative company Takeda. When it comes to domestic companies, we have RTS, MTS, Agromarket, JP Službeni glasnik [Official gazette], EPS, JP Pošta Srbije [Post of Serbia] etc.

How are Serbia’s capacities in terms of available personnel and how is interest among students when it comes to this area?

Serbia has well-educated engineers at the level of bachelor’s and master’s degrees, and there is high interest among students. I think the latest statistic I heard is that we have 49,000 programmers, most of whom have good fundamentals to deal with artificial intelligence. We unfortunately remain significantly behind the global leaders when it comes to the level of doctoral studies and research.

The mere fact that hardware intended for the development of AI solutions has been acquired, installed in the State Data Centre and made available to the research community and startups free of charge is itself truly exceptional

One important aspect of the mission of our institute is to create an environment of world-class excellence for researchers dealing with artificial intelligence in our country and to support the development of their personal excellence. Through an innovative structure for connecting our researchers with mentors around the world who have achieved a global level of excellence and introducing world-class criteria for evaluating research output, we have endeavoured to set a regional standard in this regard from the very outset, and it seems to me that we’ve succeeded.

Resources in Serbia are invested in infrastructure, hardware and software, while these same resources should be utilised by research communities, colleges and start-ups. Is this already happening and, if so, to what extent – if, for example, we compare Serbia to countries that have achieved an equal level development?

– I am perhaps biased, but I would say that – compared to other countries that have achieved an equal level of development – we are leading the way in this field. The mere fact that hardware intended for the development of AI solutions has been acquired, installed in the State Data Centre and made available to the research community and start-ups free of charge is itself truly exceptional. The software used in this field is itself usually available publicly, so it is access to adequate hardware that represents an elementary problem.

This platform has been in use for over a year. There are problems, of course, due to the need to support a large number of users in a secure way, but engaging colleagues from the Office for IT and eGovernment and the wider community in resolving them represents invaluable experience for our community.

When it comes the allocating of funds for development in the field of artificial intelligence, the only exception in the region, which I think should be mentioned, is Bulgaria, where the parliament has allocated 100 million euros for the development of its institute for artificial intelligence at the University of Sofia. We can’t even measure up to that at the moment, and nor can other countries in the neighbourhood.

You have spoken very often to the media about artificial intelligence applications and your website has a large number of written articles about AI applications. Simultaneously, your institute also organises mini-symposia to discuss AI applications in specific fields. What are the main messages that you want to convey through these forms of communication?

– To be honest, this is more a consequence of the interest this topic enjoys among the general public and calls from the media than any desire we have to convey some particular message. We try to respond to such requests as much as possible, because familiarising citizens and businesses with achievements in this area is a basic prerequisite for AI-based solutions to be accepted.