AI has been so powerful that it can predict the time of human death!
- Tram Ho
A recent AI system has been trained by scientists to collect general health information for more than half a million people in the UK, over the past decade.
They then asked the AI to predict if an individual is at risk of premature death – in other words, die earlier than the average age of a person – caused by chronic diseases.
This assessment of the AI’s ability was made by Professor Stephen Weng, who is currently working on epidemiology and data science at Nottingham University, UK. The results of anticipating AI death are judged to be “significantly more accurate” than diagnostic systems that do not use machine learning technology – large-scale information collection.
How does AI predict death?
To assess the ability of an object to die, researchers tested on two types of artificial intelligence. The first is “deep learning,” in which layered information processing networks help the system learn information from real examples.
Besides, a simpler “random forest” AI is also used to combine multiple models, according to a tree-like mind map, to consider possible results.
Finally, the scientists compared the conclusion from AI with results from a manual algorithm, called the Cox model.
The basic factors for assessment include: age, gender, smoking history, cancer screening results. While the traditional Cox model focuses on ethnicity and physical activity, the AI is interested in the percentage of body fat, consumption of vegetables, as well as the hazards from the environment such as umbrella Air, alcohol and drug abuse.
Surprising accuracy
Using these systems, scientists assessed data in the British Bio Bank (UK BioBank) – a database of genes, physics and health – formed from over 500,000 medical records. Between 2006 and 2016. During those 10 years, nearly 14,500 contributors died mainly because of cancer, heart disease and respiratory disease.
As a result, the “deep learning” apparatus gave the most accurate predictions, correctly identified 76% of dead subjects during the study period. While “random forest” gives equally impressive results with 64% accuracy. Traditional Cox model is only 44% correct on the total number of results.
Direction for AI in medicine
This is not the first time experts have harnessed the power of AI to take care of people’s health. In 2017, AI was used to diagnose early signs of Alzheimer’s with 84% accuracy.
In addition, human ability to have autism, diabetes, heart attack and stroke can also be predicted by AI.
Although it may sound odd, using this AI death determination process will “help verify the scientific methods and future development of this exciting field,” said Professor Joe Kai. , a person working for the United Nations to share.