Annotated scientific and technical report for Stage 2 under R&D Project No. 365/1796 "Development of prototype of expert decision support system based on artificial intelligence"
DOI:
https://doi.org/10.71942/w943-b524Keywords:
expert system, artificial intelligence, large language models, telemedicine, cloud service, telediagnostic platformAbstract
The relevance of the work is due to the rapid development of generative artificial intelligence (AI), in particular large language models (LLM) and their implementation in all spheres of public life. One of the key important areas of AI application is medicine and healthcare, where the accuracy and timeliness of patient diagnosis or treatment is of crucial importance. The annotated report contains a brief description of the results obtained during Stage 2 of the implementation of R&D No. 365/176 by the co-executor of the state order. Three functional patient support modules have been developed for primary (family) medicine doctors; for specialized medicine doctors; rehabilitation medicine doctors; a personal assistant module for the patient. The expert system for decision support in physical and rehabilitation medicine is developed based on a hybrid architecture and implemented in the form of a specialized chatbot MeDeBERTaBot based on the LLM Gemma 3 family. Other components include: tools for optimizing patient-centered rehabilitation programs based on cost and comfort criteria; regression models for predicting the patient's psychological state; analysis of video file processing materials containing data on the change in spatial coordinates of characteristic points over time and for measuring the linear dimensions of the human body; identification of the most dangerous ECG syndromes; processing and analysis of thermal and capillaroscopic images. The product includes program code, data sets in the form of image files, ECG recording files, system prompts, processing results are recorded in structured JSON source files. Program modules are implemented in the form of components of software libraries, accompanied by comments in the code and suitable for integration into the software environment of the customer's telerehabilitation platform. Visualization tools have been created for a visual presentation of medical data and the results of their processing and analysis.Downloads
Posted
2026-07-15
License
Copyright (c) 2026 Arxiv Academy

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.