Our team presented a paper titled “Interpretable ML for Diabetes and Prediabetes Screening Using Self-Reported Health Indicators” by S. Lazic, S. Cakic, I. Rubezic Lukic, N. Popovic, and T. Popovic at the 30. Annual Conferenc on Information Technology IT 2026. This was part of mentoring activities and efforts related to development of young researchers.
The paper was presented at the conference by Ms. Sanja Lazic (MSc candidate)
ABSTRACT – Early identification of type 2 diabetes (T2D) and prediabetes enables timely interventions, yet screening often relies on self-reported data rather than laboratory testing. This work compares lightweight Machine Learning (ML) models: Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Multilayer Perceptron (MLP) trained on 21 self-reported indicators from the 2015 Behavioral Risk Factor Surveillance System (BRFSS) dataset for three-class classification (no diabetes, prediabetes, diabetes). We propose a screening-oriented evaluation where a probability threshold is selected to achieve a target sensitivity (recall) of 0.80. LightGBM achieves balanced accuracy of 0.52 and precision of 0.33 at the target sensitivity, with 38% of cases flagged. Tree SHapley Additive exPlanations (TreeSHAP) highlight general health status, age category, body mass index (BMI), and hypertension as dominant predictors. A FastAPI web application provides individual risk estimates and instance-level explanations. The pipeline demonstrates feasibility of interpretable, calibrated screening from non-laboratory data.
AI-AGE team represenatives participated in the various events and activities at IT2026 conference in Zabljak. We ook part in discussions related to Lessons Learned for HPC and AI applications in various domains in Montenegro. We also participated in the special session dedicated to project results presentations where we had a chance to discuss the project with researchers from Montenegro and the region.
AI-AGE was featured at the project results presentations
We particiapted in EuroCC4SEE panel on Lessons Learned for HPC and AI applications
High-Performance Computing (HPC) and Artificial Intelligence (AI) are rapidly transforming the landscape of healthcare — moving far beyond research prototypes into solutions that can shape clinical practice and improve patient outcomes. This transition from strategy to real-world impact was the focus of the recent EuroCC2 initiative “Symposium: HPC and AI Driven Innovation in Healthcare”. AI-AGE team participated in organization, coordination, and presentations at this event.
Click to watch video
The Vision: Bridging Strategy and Clinical Practice
The event was organized in cross-collaboration with EuroCC2 & EuroCC4SEE
AI-AGE project was presented and further collaboration with EuroCC3 was discussed
Healthcare is today generating vast volumes of data — from medical imaging to electronic health records, genomics, wearable sensors and beyond. HPC provides the computational power needed to process and analyse this data at scale, while AI techniques such as deep learning unlock patterns that are invisible to traditional analysis methods. Together, HPC and AI form a powerful synergy for healthcare innovation:
Accelerated diagnostics: AI models trained on large annotated datasets can assist clinicians by accurately identifying disease signs in imaging and other modalities.
Biomarker discovery and precision medicine: High-throughput computing enables the discovery of subtle biological signals indicative of disease progression or treatment response.
Predictive and personalised care: HPC-enabled AI workflows can predict patient outcomes and support real-time clinical decision making.
Symposium included presentations of successs stories from the region
This strategic capability — from data to insights to impact — was the core theme explored through EuroCC2 activities in Montenegro and the wider South-Eastern Europe region.
Key Takeaways for Healthcare Innovation
From Research to Clinical Utility HPC and AI solutions are no longer confined to laboratories. With appropriate infrastructure, data governance and clinical integration pathways, these technologies are being translated into tools that support healthcare professionals in diagnosis and treatment.
Regional and Cross-Institutional Collaboration The EuroCC2 framework — including the National Competence Centre Montenegro — brings together academic institutions, healthcare providers, and technology partners to share resources, expertise, and training. These collaborative ecosystems are essential for building sustainable HPC-AI capacity in healthcare.
Capacity Building and Skills Development One of the crucial pillars of impactful HPC and AI adoption is training. Workshops, seminars, and hands-on sessions equip researchers, clinicians, and students with the skills to leverage HPC and AI tools effectively in their domains.
Enabling Infrastructure Access Through EuroCC2 and related programmes, researchers and practitioners gain access to European HPC resources — reducing barriers to entry for high-end computing and enabling complex analyses that were previously impractical.
What This Means for Montenegro and Beyond
Montenegro, alongside partner regions across Europe, is building the foundation for a healthcare ecosystem that integrates HPC and AI into everyday clinical workflows. By investing in strategic computing infrastructure, enabling cross-sector collaboration, and fostering technical expertise, the potential to improve patient outcomes, streamline clinical processes, and support data-driven medicine is growing stronger.
The event gathered representatives from Healthcare and IT sectors already involved in research and development of HPC and AI driven solutions for healthcare and medical research. More info at NCC Montenegro site: [link].
Over 20 participants in the Symposium, important discussion of next steps
High-Performance Computing (HPC) and Artificial Intelligence (AI) are increasingly moving beyond research laboratories into real clinical environments. Across Montenegro and the SEE region, promising AI solutions have been developed for medical image analysis, biomarker detection, and predictive diagnostics. The critical challenge today is ensuring their structured transition from research prototypes to validated, deployable tools within healthcare systems.
Symposium on HPC and AI in Healthcare and Medicine co-organiozed by Ai-AGE
This event addresses precisely that transition. It focuses on how HPC infrastructure, interdisciplinary collaboration, and coordinated ecosystem support can accelerate the integration of AI into everyday clinical practice. Particular attention will be given to available computational capacities, real-life use cases, and pathways toward sustainable deployment.
The event is organized as a joint initiative between NCC Montenegro and NCC Bosnia and Herzegovina, within the broader framework of EuroCC 2 and EuroCC4SEE. It also represents a form of cross-project pollination with the AI-AGE project, demonstrating how research-driven innovation can evolve into applied healthcare solutions through regional cooperation.
AI-AGE will be featured in the presentation session
Researchers, clinicians, innovators, and industry partners are invited to join the discussion, exchange expertise, and contribute to shaping the next steps for HPC- and AI-driven healthcare across Southeast Europe. The event is scheduled for Friday, 13 Feb 2026. Please contact NCC Montenegro for further details.
We are proud to share a newly published paper co-authored by an AI-AGE researcher within the Horizon 2020 RECOGNISED consortium in DIABETOLOGIA. This research resulted from collaboration of leading European experts in diabetes, diabetic retinopathy and cognitive impairment. The study shows that retinal neurodysfunction is linked to mild cognitive impairment in people with type 2 diabetes, reinforcing the concept of the retina as a window into brain health.
Beyond its scientific findings, the project marks an important step in strengthening international research infrastructure. Our team helped implement rigorous Good Clinical Practice standards and harmonised imaging and cognitive assessment protocols across multiple European centres — groundwork essential for future large, high-quality datasets suitable for AI-driven discovery of early, non-invasive biomarkers of aging and multimorbidity. This collaboration advances the AI-AGE mission to expand cross-border research capacity and supports ethically robust, standardised big-data research in aging and diabetes. Link to paper: https://doi.org/10.1007/s00125-025-06664-4
The AI-AGE project will be actively featured at the INFORMATION TECHNOLOGIES 2026 (IT 2026), the 30th edition of a long-standing international scientific and professional conference bringing together academia, industry, and the wider innovation ecosystem. The conference taktes place in Zabljak, 24-28 February.
As part of the project presentation session, AI-AGE will showcase its ongoing research on AI-driven biomarkers of ageing, with a particular focus on advanced data analytics, computer vision, and high-performance computing applied to biomedical and health-related domains. This session provides an excellent opportunity to present project objectives, methodologies, and early scientific insights to a broad and interdisciplinary audience.
Click on image to open IT2026 website
In addition to the project presentation, AI-AGE researchers will actively participate in the conference program, contributing to scientific discussions and knowledge exchange across multiple sessions. Importantly, young researchers involved in the AI-AGE project will present a scientific paper, highlighting the project’s strong commitment to capacity building, early-stage researcher development, and excellence in applied AI research.
The FFplus Innovation Studies Open Call is now open, offering European SMEs and start-ups a focused opportunity to develop and validate generative AI (GenAI) solutions using large-scale European supercomputing resources. The call opens on 3 February 2026 and targets early-stage, high-impact projects with a clear proof-of-concept orientation.
Click to open application website
Coordinated by HLRS (High-Performance Computing Center Stuttgart), FFplus supports innovation at the intersection of AI and HPC, enabling companies to overcome computational barriers and scale beyond conventional cloud or on-premise infrastructures.
Feel free to check with NCC Montenegro for help with the application.
Key facts
Funding: up to €300,000 per project
Deadline: 25 February 2026 (or earlier if 250 proposals are submitted)
In line with the AI-AGE project’s commitment to open science and transparency, the dataset associated with our newly published BMJ Open paper has been made publicly available on Zenodo.
The dataset includes anonymised primary care data on frailty indicators, chronic diseases, and demographic variables among adults aged 40–69 in Montenegro. It provides a valuable resource for researchers working on biological ageing, frailty, multimorbidity, and AI-driven risk modelling, particularly in middle-income and transitional health systems.
By sharing this dataset, AI-AGE aims to support international collaboration, reproducibility, and secondary analyses that can further advance non-invasive ageing research.
We are pleased to announce the publication of the AI-AGE project’s first peer-reviewed paper in BMJ Open: “Frailty and multimorbidity among adults aged 40–69 years in Montenegro: a cross-sectional pilot study.”
This is the first study to provide population-based evidence on frailty, prefrailty, and multimorbidity in middle-aged adults in Montenegro. The key finding with direct relevance for Montenegro is that more than half of adults aged 40–69 already show signs of prefrailty or frailty, with a sharp increase in chronic disease burden observed from the age of 55 onwards. This highlights a critical window for early prevention and intervention within primary healthcare, well before old age.
The study also demonstrates that frailty screening is feasible in routine primary care settings, providing a strong foundation for scaling up AI-supported, non-invasive biomarkers of ageing within the Montenegrin healthcare system.
🔗 Read the full open-access paper at the following [link].
AI-Age project is featured in promo videos created by the Ministry for Education, Science and Innovation. Our PhD students dr Isidora Rubezic-Lukic and mr Dejan Babic presentd the progress on the project and current results (MPNI on Instagram link).
Developing different AI tools to support identification of new biomarkers of aging
The team successfully established all core foundations for research, including launching the official project website, extending UK Biobank data access to both partner institutions, and deploying new HPC/AI infrastructure needed for biomedical image analysis. Young researchers received extensive training through EUROCC, TRACEWINDU, HPC4S3ME, and AIFusion programs, while interdisciplinary cooperation between UDG and the Faculty of Medicine was strengthened. The project also achieved early scientific dissemination through a poster presentation at EAsDEC 2024 in Milan and participation at IT2024 in Žabljak.
mr Dejan Babic (PhD candidate) in the AI lab at the Faculty for information systems and technologies
In parallel, significant progress was made on doctoral research activities: one initial PhD proposal was submitted and defended, and research questions for two doctoral theses were defined in alignment with AI-AGE goals. WP1 and WP2 outputs were fully completed, while WP3 research—focused on AI/ML models for retinal biomarkers of aging—was initiated ahead of schedule. These accomplishments provide a strong foundation for validating hypotheses, developing advanced AI models, and expanding international collaboration in the second project year.
Promo video on AI-AGE created and posted by the Ministry (MPNI)
