K. Zarkogianni

Konstantia Zarkogianni received a MEng in Electrical and Computer Engineering (2003) from the Aristotle University of Thessaloniki, a MSc Degree in Electronic and Computer Engineering (2005) from the Technical University of Crete, and a PhD degree (2011) from the National Technical University of Athens (NTUA), Greece. In October 2017, she was appointed as a permanent laboratory teaching staff member at the School of Electrical and Computer Engineering of the NTUA. Since 2023, she has been an Associate Professor of Human-Centered AI at DACS of UM. Her research is mainly focused on AI (e.g., ML, XAI, knowledge bases & reasoning, and recommender systems) and decision intelligence. She has authored or coauthored 18 papers in refereed international journals, one chapter in book, and more than 30 papers in international conference proceedings. She has participated as research associate and PI in greek and EU funded projects. She has been a guest editor of the special issue on Emerging Technologies for the Management of Diabetes Mellitus (Springer Journal of Medical and Biological Engineering and Computing [MBEC], 2015). She has been a member of the Editorial Board of the SpringerPlus journal in 2016 and reviewer for international scientific journals (IEEE Transactions on Biomedical Engineering, IEEE Journal of Biomedical and Health Informatics, Springer Medical & Biological Engineering & Computing, Elsevier Journal of Biomedical Informatics, and JSM Diabetology and Management). She is a member of the Institute of Electrical and Electronics Engineers (ΙΕΕΕ) and the Technical Chamber of Greece.

Expertises
  • Intelligent decision support systems: Design, development and evaluation of decision support systems able to produce personalized and adaptive predictions/recommendations. The development of the systems is based on AI technologies such as e.g. machine learning, knowledge representation and reasoning, and recommender systems. Particular interest in addressing the important challenge of mitigating biases that might occur within each stage in the AI-aided decision-making process: dataset (e.g. imbalances, data shifts, limited size, and missing information), model (e.g. systematic errors), and human (e.g. cognitive).
  • Biomedical and health informatics: Efficient management and processing of multi-level and multi-scale biomedical data gathered from various data sources (e.g. sensors, wearables, Electronic Health Records, medical imaging etc). Interdisciplinary approach that combines a priori medical knowledge and current clinical guidelines with cutting edge AI technologies towards delivering P4 solutions (preventive, predictive, personalized, and participatory) for chronic disease management.    
  • Explainable AI: Development of innovative explainable and interpretable techniques able to generate reliable explanations at the attribution, semantic and examples level. At the attribution level, particular focus is placed on realizing robustness against randomly generated perturbations (e.g. LIME techniques) while exploiting unrolling methods to explain recurrent neural networks. At the semantic level, the use of neural-symbolic learning techniques is explored towards producing contrastive, selective, interactive, social, and situational explanations for end-users that are not AI-experts. At the examples level, the use of Generative AI is being investigated.      
  • Human in the loop technologies: Design and implementation of Hybrid Intelligence that leverages the strengths of human thinking and the AI’s capacity to perform complex calculations. In depth research in the emerging technologies of active learning, machine teaching and interactive machine learning in order to achieve effective interaction between human and machine intelligence.