Price: 0
Number of applications: 7
04.12.25 (inclusive)
under discussion
Idea
ICT tasks
Robotics
Intelligent control systems
Software/ IS
The system should accept data from the following sources: patient complaints (text/structured input); medical history (medical history, heredity, lifestyle); test results (blood, urine, biochemistry, etc.); results of instrumental studies (ultrasound, MRI, CT — in the form of text protocols); physiological data (blood pressure, pulse, temperature, etc.); data from medical devices/sensors (optional). Natural language processing (NLP) recognition and normalization of medical terms; identification of symptoms, signs, risk factors; building a structured text representation. Diagnostic Module (AI module) The system should: create a list of probable diagnoses indicating the probability; identify dangerous conditions and issue alert messages (for example, suspected stroke, heart attack, sepsis); form the rationale for hypotheses (explainability): "diagnosis X is assumed based on symptoms A, B, C and the results of analysis D".
The clinical recommendations module provides recommendations for necessary additional examinations; drug compatibility verification; warnings about contraindications; verification of compliance with clinical protocols (based on local/international guidelines — if licensed). Doctor's interface view input data; view diagnostic hypotheses of AI; mark "confirmed/refuted"; formation of the final diagnosis by the doctor; decision log (audit log).
Turdieva Dina Turarovna
Purpose and description of task (project)
To develop a medical assistant (hereinafter referred to as the “System”) who analyzes the patient's clinical data and forms probabilistic diagnostic hypotheses, recommendations for the doctor and warnings about risks. The system should minimize the likelihood of diagnostic errors and comply with quality standards and medical regulations.
Whatsapp
