The international scientific community presented a new technology of revolutionary significance for the medical field. Scientists have developed an artificial intelligence model that can predict the risks of various diseases that may arise in human health in the long term. It is noted that this system, called Delphi-2M, can analyze more than a thousand types of diseases and sometimes show problems 10 years in advance.
The operating principle of the model is implemented by combining doctor's records, data collected in the hospital, and factors of a person's lifestyle (for example, smoking habits, diet, level of physical activity). Experts note that Delphi-2M displays health risks based on probabilistic conclusions, such as weather forecasts.
This artificial intelligence was initially trained in UK Biobank data (more than 400,000 participants) in Great Britain. Subsequently, data from more than 1.9 million citizens in Denmark's national healthcare system were examined. The results of the study showed that Delphi-2M with high efficiency detects such dangerous diseases as heart attack, diabetes, and sepsis. However, it was noted that its capabilities in predicting accidental infections are still limited.
Scientists believe that this technology could radically change medical practice in the future. With the help of artificial intelligence, doctors can provide individual consultations to patients, recommend lifestyle changes to reduce the risk, or take preventive treatment measures if necessary. It is also of great importance for the entire healthcare system, as it allows for predicting the trends in the spread of diseases and planning the workload of hospitals.
The project was implemented in cooperation with scientists from the German Center for Cancer Research (DKFZ), the European Laboratory of Molecular Biology, and the University of Copenhagen. The research results were published in the prestigious international scientific journal Nature.
In the future, specialists plan to add a database of genetic data, blood tests, and medical images to the model for further improvement. At the same time, it is planned to work with even larger clinical data covering different age groups.
Read 'Zamin' on Telegram!
