About
Key areas:
- Development of reliable and reproducible AI methods
- Explainable AI (XAI)
- Multimodal AI
- Agentic / agent-based AI methods
Vision:
To become an AI methods research lab that is integrated into the European Research Area and internationally visible, contributing to the development of European AI solutions and models.
Mission:
To develop reliable, transparent and reproducible AI methods; strengthen PhD and postdoctoral competencies; integrate advanced research into studies; and build an ecosystem through projects, workshops, summer schools and national initiatives.
Goal:
To develop reliable, transparent and reproducible AI methods and the experimentation infrastructure for them, ensuring robustness and reproducibility of results.
Objectives:
- Develop and advance explanatory methods that increase model transparency and explainability.
- Develop multimodal AI methods (e.g., remote sensing, sensor data, language).
- Advance agentic AI methods for automated, explainable evaluation and orchestration.
- Ensure reproducibility through experimental design, metrics, benchmarks and testing.
- Develop talent (PhD students, postdoctoral researchers, researchers, engineers) and expand international collaboration (e.g., EU consortia).
Infrastructure:
AIML leverages VU MIF computing resources, including HPC infrastructure enabling large-scale training, evaluation and data-intensive research.
Collaboration & opportunities:
- R&D projects (fundamental and applied): methods → real-world impact
- Partnerships in international calls (Horizon Europe, etc.)
- Collaboration with startups and industry (pilots, technology transfer, joint IP)
- Talent attraction: PhD candidates, postdocs, research engineers, visiting researchers
Contact:
Head of Lab: Assoc. Prof. Dr. Valentas Gružauskas
Email:
Phone: +370 685 76870
Publications
2025
- Multispectral image caption unification using diffusion and cycle GAN models, Komurcu, K., Petkevičius, L., IEEE Access, DOI: 10.1109/ACCESS.2025.3632152
- MiniCPM-V LLaMA model for image recognition: a case study on satellite datasets, Kömürcü, K., Petkevičius, L., IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, DOI: 10.1109/JSTARS.2025.3547144
- Designing a framework for ethnography-driven prompt engineering in social work, Seniutis, M., Sas, A., Gružauskas, V., Navickas, V., Švažas, M., Human Technology, DOI: 10.14254/1795-6889.2025.21-1.5
- Dynamic changes in depressive symptoms at the onset of military conflict in a neighboring country: a cross-sectional study, Airapetian, A., Gružauskas, V., Urbonaitė, N., Bachmetjev, B., Bernadickas, P., Nedzinskienė, L., Zablockis, R., Frontiers in Psychiatry, DOI: 10.3389/fpsyt.2025.1593088
- Few-shot learning for triplet-based EV energy consumption estimation, Čivilis, A., Petkevičius, L., Šaltenis, S., Torp, K., Markucevičiūtė-Vinckė, I., Applied Artificial Intelligence, DOI: 10.1080/08839514.2025.2474785
- Optimising quantum algorithm schemes according to the topology of quantum computers, Keibas, G., Petkevičius, L., 12th IEEE Workshop on Advances in Information, Electronic and Electrical Engineering (AIEEE), DOI: 10.1109/AIEEE66149.2025.11050773
- Reproducibility of Ki67 Haralick entropy as a prognostic marker in estrogen receptor-positive HER2-negative breast cancer, Žilėnaitė-Petrulaitienė, D., Rasmusson, A., Valkiūnienė, R. B., Laurinavičienė, A., Petkevičius, L., Laurinavičius, A., American Journal of Clinical Pathology, DOI: 10.1093/ajcp/aqaf081
- Intratumoral heterogeneity of Ki67 proliferation index outperforms conventional immunohistochemistry prognostic factors in estrogen receptor-positive HER2-negative breast cancer, Žilėnaitė-Petrulaitienė, D., Rasmusson, A., Besusparis, J., Valkiūnienė, R. B., Augulis, R., Laurinavičienė, A., Plancoulaine, B., Petkevičius, L., Laurinavičius, A., Virchows Archiv, DOI: 10.1007/s00428-024-03737-4
- Prostate cancer identification from pathology diagnosis descriptions in Lithuanian language, Griguola, T., Smailytė, G., Petkevičius, L., 12th IEEE Workshop on Advances in Information, Electronic and Electrical Engineering (AIEEE), DOI: 10.1109/AIEEE66149.2025.11050774
Team
Valentas Gruzauskas, Head of Lab
Linas Petkevičius, Chief researcher
Dr. Vytautas Valaitis, Scientist
Boleslovas Dapkūnas, Juniour Researcher
Partn. doc. Jonas Matuzas, Partn Prof.
Justinas Lekavičius, Juniour Researcher
Doc. Linas Bevainis, Assoc. Prof.
Dr. Vytautas Paura, Scientist
Lect. Vytenis Šliogeris, Lector
Marius Šomka, Specialist
Evgenij Shapovalov, Specialist
Artiom Hovhannisyan, Specialist
Augustė Raišytė, Specialist
Kursat Kömürcü, Specialist
Povilas Kvedaras, Specialist
Saulė Satkauskienė, Specialist
