Periodic Reporting for period 1 - GAIN (Georgian Artificial Intelligence Networking and Twinning Initiative)
Période du rapport: 2022-10-01 au 2025-09-30
The overall goal of the project was to integrate MICM into the European research community in AI by strengthening its research and innovation capacities aiming to create a European-level Centre of Excellence in this field in Georgia and the South Caucasus region. The leading European research organizations supported MICM in this endeavour: DFKI, Germany, INRIA, France, and EXOLAUNCH GmbH, Germany.
The project pursued the following operational objectives:
Objective 1. To implement the Twinning Research and Innovation Programme
Objective 2. To implement the targeted Scientific Excellence and Networking Capacity Building Programme.
Objective 3. Investing in young talents.
Objective 4. To build a new quality of innovation and research management skills for the MICM research staff.
Objective 5. To improve the visibility and image of MICM.
Objective 6. To design the measures ensuring the sustainability and organic development of the project results.
Strategic Impact of GAIN:
• Strengthening Research Capacity – Recruitment and training of young researchers
• Integration into the European Research Area in AI – Active collaboration with INRIA and DFKI.
• Sustainability and Institutional Development – Best practices in research administration and management inspired the creation of a new structural unit, the MICM AI Department
• Enhanced Scientific Reputation – MICM was funded for 4 new research projects, including coordination of FORGE-AI.
• Visibility – Organization of workshops and summer schools, including the SCCAI 2025 Final Conference.
• Long-Term Collaboration – Foundations were laid for continued joint research with leading European partners.
The project pursued the following operational objectives:
Objective 1. To implement the Twinning Research and Innovation Programme
Objective 2. To implement the targeted Scientific Excellence and Networking Capacity Building Programme.
Objective 3. Investing in young talents.
Objective 4. To build a new quality of innovation and research management skills for the MICM research staff.
Objective 5. To improve the visibility and image of MICM.
Objective 6. To design the measures ensuring the sustainability and organic development of the project results.
Strategic Impact of GAIN:
• Strengthening Research Capacity – Recruitment and training of young researchers
• Integration into the European Research Area in AI – Active collaboration with INRIA and DFKI.
• Sustainability and Institutional Development – Best practices in research administration and management inspired the creation of a new structural unit, the MICM AI Department
• Enhanced Scientific Reputation – MICM was funded for 4 new research projects, including coordination of FORGE-AI.
• Visibility – Organization of workshops and summer schools, including the SCCAI 2025 Final Conference.
• Long-Term Collaboration – Foundations were laid for continued joint research with leading European partners.
The basis for all research capacity building activities of the GAIN project was the Twinning Research and Innovation Programme (TRIP). The programme was concentrated around 3 research topics: AI Methods for Deep Speech Analysis in Health, AI Technologies for Human Behaviour Understanding, Pilot Research Project at MICM, based on the INRIA-DFKI large-scale project MEPHESTO. Georgian researchers were assigned to the specific research topics, sub-tasks supervised by INRIA and DFKI.
To lay the groundwork for the successful implementation of the project, the first step was the recruitment of young researchers. Among 27 Georgian researchers of the GAIN-MICM team, incorporated in the “MICM AI Lab”, 18 were PhD, Master and Bachelor students from MICM, GTU, other universities, institutions and 2 AI start-ups. From DFKI and INRIA, 6 researchers and 5 PhD students were involved in the joint research. 4 students from MICM held the Tenure-track Positions who leaded 4 research groups and participated in the Staff Exchange Scheme to visit DFKI and INRIA for about 3 months. MICM created 5 new employee positions which are held by the GAIN students. The “Joint Virtual Laboratory” was composed to continue cooperation beyond the GAIN project.
During the project implementation, 27 joint research papers were published in peer reviewed, open access journals, 2 papers were accepted for publication, 1 is submitted and 1 prepared for submission. 39 conference talks were delivered.
Research collaboration resulted as well in open-source software products: 4 datasets, 5 software and 3 models.
The project organized 23 research and training events: two GAIN Summer Schools; 3 GAIN Workshops at ACII 2024, IPAS25 and RANLP-2025; the Final Conference SCCAI2025, where 5 keynote speakers were invited from Armenia, Azerbaijan, France and Germany, and 3 Georgian students were the co-chairs of the conference tracks.
The trainings for researchers, administrators and managers were delivered on topics such as Personnel Management and Motivation, Scientific Fundraising, etc. Thus, the project has supported the organizational changes in MICM. As such, in the end of 2025, MICM AI Lab will be transformed to the research unit “MICM AI Department”.
During the project’s lifetime, 7 research project proposals involving MICM were submitted, of which 4 (2 international) were funded. MICM coordinates one of the largest HORIZON EUROPE AI projects in Georgia, FORGE-AI (GA #101216703). MICM is invited as a partner in the HE proposal WARIE (Proposal number: 101311127) demonstrating the institute’s growing recognition.
To lay the groundwork for the successful implementation of the project, the first step was the recruitment of young researchers. Among 27 Georgian researchers of the GAIN-MICM team, incorporated in the “MICM AI Lab”, 18 were PhD, Master and Bachelor students from MICM, GTU, other universities, institutions and 2 AI start-ups. From DFKI and INRIA, 6 researchers and 5 PhD students were involved in the joint research. 4 students from MICM held the Tenure-track Positions who leaded 4 research groups and participated in the Staff Exchange Scheme to visit DFKI and INRIA for about 3 months. MICM created 5 new employee positions which are held by the GAIN students. The “Joint Virtual Laboratory” was composed to continue cooperation beyond the GAIN project.
During the project implementation, 27 joint research papers were published in peer reviewed, open access journals, 2 papers were accepted for publication, 1 is submitted and 1 prepared for submission. 39 conference talks were delivered.
Research collaboration resulted as well in open-source software products: 4 datasets, 5 software and 3 models.
The project organized 23 research and training events: two GAIN Summer Schools; 3 GAIN Workshops at ACII 2024, IPAS25 and RANLP-2025; the Final Conference SCCAI2025, where 5 keynote speakers were invited from Armenia, Azerbaijan, France and Germany, and 3 Georgian students were the co-chairs of the conference tracks.
The trainings for researchers, administrators and managers were delivered on topics such as Personnel Management and Motivation, Scientific Fundraising, etc. Thus, the project has supported the organizational changes in MICM. As such, in the end of 2025, MICM AI Lab will be transformed to the research unit “MICM AI Department”.
During the project’s lifetime, 7 research project proposals involving MICM were submitted, of which 4 (2 international) were funded. MICM coordinates one of the largest HORIZON EUROPE AI projects in Georgia, FORGE-AI (GA #101216703). MICM is invited as a partner in the HE proposal WARIE (Proposal number: 101311127) demonstrating the institute’s growing recognition.
Open-source software products:
1. EEGain - designed to improve the comparability and generalizability of EEG emotion recognition approaches
2. EmotionVMAE - designed to improve emotion recognition from video
3. Tokenization of the Georgian Language - designed to improve Georgian language models
4. LLM for Low Resourced Languages. This model, designed jointly with the project partner “AI Lab” is an analogue e.g. of ChatGPT for the Georgian language
The first 3 products, though innovative, require further research and development efforts to advance toward commercialization and market readiness.
Model #4, by contrast, demonstrates high scientific, societal, technological, and economic potential, and has already reached TRL 7, with a prototype successfully demonstrated in an operational environment. The project has implemented a set of innovation activities to bring it closer to the market, including business case development and business modelling, market research, IP regime definition, initial marketing, etc.
1. EEGain - designed to improve the comparability and generalizability of EEG emotion recognition approaches
2. EmotionVMAE - designed to improve emotion recognition from video
3. Tokenization of the Georgian Language - designed to improve Georgian language models
4. LLM for Low Resourced Languages. This model, designed jointly with the project partner “AI Lab” is an analogue e.g. of ChatGPT for the Georgian language
The first 3 products, though innovative, require further research and development efforts to advance toward commercialization and market readiness.
Model #4, by contrast, demonstrates high scientific, societal, technological, and economic potential, and has already reached TRL 7, with a prototype successfully demonstrated in an operational environment. The project has implemented a set of innovation activities to bring it closer to the market, including business case development and business modelling, market research, IP regime definition, initial marketing, etc.