[ad_1]
Artificial intelligence is already upon us, improving our way of life in almost every aspect. Even in its infancy, AI could power personal badistants such as Alexa and Siri and even control prototypes of autonomous cars. Researchers have however revealed that a new generation of artificial intelligence algorithms could already "very accurately" predict when those struggling against prolonged living conditions will die.
The medical records of half a million people have been developed to develop an AI tool to predict who is at risk of premature death.
The data takes into account everything from the patient's medical history regarding disease, salt intake, use of medication and use of sunscreen.
Researchers at the University of Nottingham described the machine learning system as "very accurate" in testing – about 10% more reliable than estimates of existing AI systems.
The groundbreaking research was led by Dr. Stephen Weng, an badistant professor of epidemiology and data science at the university.
Dr. Weng said, "Preventive health care is a growing priority in the fight against serious diseases.
READ MORE: The mobile nuclear reactor of the US military is "a colossal mistake"
"We have been working for several years to improve the accuracy of the computerized badessment of health risks in the general population.
"Most applications focus on one disease area, but the prediction of death from a number of different diseases is extremely complex, particularly because of environmental and individual factors that may affect it.
"We have taken a big step forward in this area by developing a unique and holistic approach to predicting the risk of premature death of a person through machine learning."
The artificial intelligence used data from 502,628 adults of average age between 2006 and 2010, and was followed for a decade.
In total, 60 predictors of health were taken into account, from the body mbad index (BMI) of participants to blood pressure and vitamin intake.
READ MORE: The NASA photo shows the deepest view of the universe of all time
And the algorithm also included information about the diet and alcohol consumption of the participants.
Dr. Weng added, "We mapped the resulting predictions with mortality data from the cohort with the help of death records from the Office of National Statistics, the UK Cancer Registry and statistics on "hospital episodes". "
The generated model was then contrasted with two standard machine learning techniques.
Published results indicate that the latest RN model was 10.1% more accurate than the existing AI technology.
"We found that machine-learned algorithms were much more accurate in predicting death than standard prediction models developed by a human expert," said Dr. Weng.
READ MORE: Hunting for foreigners "just getting started" says high-level scientist
Professor Joe Kai, author of the studies, responsible for primary care at the School of Medicine and Health Sciences, added: "The potential to use" AI "or" learning by machine "arouses a keen interest in better predicting health outcomes.
"In some situations, it may help, in others, it is not necessarily the case.
"In this particular case, we have shown that with careful tuning, these algorithms can usefully improve the prediction.
"These techniques can be new to many health researchers and difficult to follow.
"We believe that by clearly declaring these methods in a transparent manner, this could help in the scientific audit and future development of this exciting area for health care."
READ MORE: Has NASA lied about the landing of the moon?
Source link
