Stabilarity Research Hub: Explainable AI for Medicine, Economics & Tech

Who We Are

Stabilarity Research Hub is an experimental research collective formed by scientists from Odesa National Polytechnic University (ONPU), Ukraine. Our researchers come from two interdisciplinary departments: Economic Cybernetics and Artificial Intelligence/Machine Learning.

This unique combination gives us a distinctive perspective — we study AI not just as a technical challenge, but through the lens of economic viability, real-world implementation costs, and societal impact.

Our Focus: Explainable AI

We believe AI systems should be transparent and interpretable. Our core research focuses on Explainable AI (XAI) and its influence across three key industries:

Healthcare — Medical imaging diagnosis with visual explanations
Economics — Cost-benefit analysis of AI implementations
Technology — Enterprise AI architecture and deployment patterns

The Stabilarity Research Hub

Our hub publishes open research on AI economics, medical ML, and emerging technologies. All articles are peer-reviewed and registered with DOI on Zenodo. We maintain complete transparency about our methods and limitations.

→ Visit Stabilarity Research Hub

Medical ML Diagnosis Research — Complete

Our flagship project: a 43-article research series on machine learning for medical imaging diagnosis, authored by Oleh Ivchenko with contributions from Dmytro Grybeniuk.

The research covers:

CNN and Vision Transformer architectures for medical imaging
Explainable AI techniques (Grad-CAM, attention visualization)
Clinical workflow integration strategies
Adaptation frameworks for Ukrainian healthcare systems
Cost-benefit analysis for hospital administrators

→ Full Medical ML Research Series

ScanLab — From Research to Practice

The Medical ML research culminated in ScanLab — an open-source medical imaging diagnostic platform. Built on PyTorch and FastAPI, it features:

Binary classification with probability scores
Grad-CAM visualizations showing which image regions influenced the prediction
DICOM support for clinical integration
Ukrainian/English bilingual interface
Batch processing for high-volume facilities
Analytics endpoints for ROI tracking

Note: ScanLab is an experimental research tool. It is not a certified medical device and should not be used for clinical decision-making without proper validation.

What We Are — And What We Are Not

We are an experimental research hub. We publish exploratory work, test hypotheses, and share our findings openly. We do not claim production-ready accuracy or clinical certification.

Our goal is to support AI researchers worldwide by providing:

Open research with DOI-registered publications
Practical implementation frameworks
Economic analysis tools for AI adoption decisions
Working prototypes (like ScanLab) for experimentation

Explore Our Research

Published by Stabilarity Research Hub — Scientists from Odesa National Polytechnic University, Ukraine 🇺🇦