Neural network graph architecture of transparent artificial intelligence in medicine
https://doi.org/10.25881/18110193_2025_2_70
Abstract
The paper presents an approach to design an information system based on a neural network graph architecture. This approach is designed to mitigate the problem of explicit explanation of decisions made by artificial intelligence — the problem of transparency (explainability, reliability, trustworthiness). The use of artificial intelligence technologies in medicine has a “transversal” character and contributes to the creation of conditions for improving efficiency and formation of fundamentally new areas of activity: automation of routine (repetitive) operations; use of autonomous intelligent equipment and robotic complexes, intelligent control systems; increasing the efficiency of planning, forecasting and medical decision-making processes. A promising technology of the proposed approach is the use of graph neural network architecture as part of the information system for data processing and analysis. In this article we introduce an example of graph node classification on an open dataset with cardio-data of conditionally healthy people and patients.
About the Authors
D. A. Andrikov
Bauman Moscow State Technical University
Russian Federation
Russian Federation
PhD, Associate Professor
Moscow
Dm. A. Andrikov
Engineering Academy, RUDN University
Russian Federation
Russian Federation
Ph.D., Associate Professor
Moscow
D. V. Berezkin
Bauman Moscow State Technical University
Russian Federation
Russian Federation
PhD, Associate Professor
Moscow
A. Iu. Popov
Bauman Moscow State Technical University
Russian Federation
Russian Federation
DSc, Associate Professor
Moscow
A. V. Proletarsky
Bauman Moscow State Technical University
Russian Federation
Russian Federation
DSc, Professor
Moscow
References
1. Karpov OE, Andrikov DA, Maksimenko VA, Hramov AE. Prozrachnyĭ iskusstvennyĭ intellekt dlya mediciny. Vrach i informacionnye tekhnologii. 2022; 2: 4-11. (In Russ.)
2. Karpov OE, Hramov AE. Prognosticheskaya medicina. Vrach i informacionnye tekhnologii. 2021; 3: 20. (In Russ.)
3. Sharova DE, Mihajlova AA, Gusev AV, Garbuk SV, Vladzimirskij AV, Vasil'ev YUA. Analiz mirovogo opyta v regulirovanii ispol'zovaniya medicinskih dannyh dlya celej sozdaniya sistem iskusstvennogo intellekta na osnove mashinnogo obucheniya. Vrach i informacionnye tekhnologii. 2022; 4: 28-39. (In Russ.)
4. Karpov O. E., Hramov A.E. Informacionnye tekhnologii, vychislitel'nye sistemy i iskusstvennyĭ intellekt v medicine. M.: DPK Press, 2022. (In Russ.)
5. Vasil'ev YuA, et al. Podgotovka nabora dannyh dlya obucheniya i testirovaniya programmnogo obespecheniya na osnove tekhnologii iskusstvennogo intellekta : Uchebnoe posobie. М.: Izdatel'skie resheniya, 2024. 140 р. (In Russ.)
6. Khorev VS, et al. Disruptions in segregation mechanisms in fmri-based brain functional network predict the major depressive disorder condition. Chaos, Solitons & Fractals. 2024; 188: 115566.
7. Kuchin AS, Grubov VV, Maksimenko VA, Utyashev NP. Avtomatizirovannoe rabochee mesto vracha epileptologa s vozmozhnost'yu avtomaticheskogo poiska pristupov epilepsii. Vrach i informacionnye tekhnologii. 2021; 3: 62-73. (In Russ.)
8. Grubov VV, Nazarikov SI, Kurkin SA, Utyashev NP, et al. Two-stage approach with combination of outlier detection method and deep learning enhances automatic epileptic seizure detection. IEEE Access. 2024; 12: 122168-122182.
9. Wang J, Yang J, Deng J, Gunes H, Song S. Graph in Graph Neural Network. 2024. doi: 10.48550/arXiv.2407.00696.
10. Goldberger A, Amaral L, Glas, L, Hausdorff J. PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation [Online]. 2000; 101(23): e215-e220.
11. Gow B, Pollard T, Nathanson LA, Johnson A, Horng S. MIMIC-IV-ECG: Diagnostic Electrocardiogram Matched Subset (version 1.0). PhysioNet. 2023. doi: 10.13026/4nqg-sb35.
12. Zhong M, Li F, Chen W. Automatic arrhythmia detection with multi-lead ECG signals based on heterogeneous graph attention networks. Mathematical biosciences and engineering: MBE. 2022; 12(19): 12448-12471.
13. Brody S, Alon U, Yahav E. How Attentive are Graph Attention Networks?. 2021. doi: 10.48550/arXiv.2105.14491.
14. Kamyar Z, Marco G. Graph Neural Networks for Topological Feature Extraction in ECG Classification. Springer Nature Singapore. Singapore, 2023. P.17-27.
Review
For citations:
Andrikov D.A., Andrikov D.A., Berezkin D.V., Popov A.I., Proletarsky A.V. Neural network graph architecture of transparent artificial intelligence in medicine. Medical Doctor and Information Technologies. 2025;(2):70-83. (In Russ.) https://doi.org/10.25881/18110193_2025_2_70
Views:
132
Email this article 