EU-Georgia collaboration is a win-win for AI research and innovation

As Georgia looks to build its artificial intelligence (AI) capacities, what better place to start than where its first computer was deployed?

"Founded in 1956, the Muskhelishvili Institute of Computational Mathematics (MICM) has long been at the forefront of computing, programming and applied research," says MICM deputy director George Giorgobiani. "But to maintain this leadership position, MICM must embrace emerging technologies like AI - a process that starts with ensuring our staff have the right skill set."

This is where the GAIN project came into play.

A unique collaboration between Georgia and the EU, the project set out to position MICM as an AI hub, one capable of attracting the country's best and brightest talent.

"By partnering with leading European Centres of Excellence DFKI [German Research Center for Artificial Intelligence] and Inria [France's National Institute for Research in Digital Science and Technology], as well as with a high-tech company, we substantially advanced our AI research capacities," adds Giorgobiani, who coordinated the project.

Creating new opportunities in AI

With a focus on early career support, the project trained 27 Georgian researchers, including young scientists and PhD, master and bachelor students. Training was primarily provided via research webinars, workshops, innovation, on-the-job training opportunities, and 'summer school' programmes.

"Taking place in both France and Georgia, these summer schools were designed to enhance scientific excellence, build capacity and train young researchers - ultimately improving their networking skills within an international setting," explains Giorgobiani.

However, the project was not only about teaching. It also provided these new researchers with ample opportunities to put their new skills to work. This included equipping MICM with such cutting-edge technologies as a high-performance computing server that researchers could use to run their AI algorithms.

Using MICM's new state-of-the-art AI Lab, researchers were also able to take part in a pilot innovation project that saw them develop four datasets, five open access software programmes and two AI models.

"The GAIN project generated significant new opportunities for the advancement of AI research and innovation in Georgia - right at the time the country is developing a national AI strategy and looking to strengthen its European connections," notes Giorgobiani.

Building a unique AI niche

But it isn't just Georgian researchers who benefit from the GAIN project. "Our EU partners and the European AI community in general all gain from having access to MICM's unique niche of expertise," he says.

According to Giorgobiani, that niche includes the use of AI in medical applications and language technologies.

As to the former, with the project's support, MICM developed tools for the early detection and monitoring of neurodegenerative diseases and mental health disorders. It also designed AI-driven systems capable of supporting psychiatric medical records, as well as electroencephalogram datasets for neuropsychological research.

On the language front, MICM helped a local start-up develop several large language models (LLMs) for the Georgian language and move them towards commercialisation. "The project effectively supported the transfer of developed technology from the academic environment to the economic sector, thereby contributing to the country's broader innovation ecosystem," remarks Giorgobiani.

The LLMs could be of interest to European businesses looking to break into the local market. They could also play an important role in facilitating EU-Georgia communications and cooperation, especially around low-resourced languages.

An ever-expanding impact

While the GAIN project has accomplished its mission of promoting and integrating MICM into the European AI research community, its impact is far from over.

Having established a permanent AI department, MICM is now fully committed to expanding its AI footprint. It has already created new collaborations with other universities and research organisations, including some located in Armenia, Azerbaijan and Ukraine, while also building on the collaborations established with its EU partners.

The project further laid the foundation for additional EU-Georgia collaborations, such as the EU-funded (FORGE-AI) project and WARIE project currently under proposal.

The project presented its many achievements at the South Caucasus Conference on Artificial Intelligence. "The cross-border collaborations we established during the GAIN project not only position MICM as a regional powerhouse in AI research and development, they play a critical role in developing the technological prowess of an EU candidate country," Giorgobiani concludes.