A new generation of AI large-scale models for predicting the time of human death

According to a report by New Scientist, a research team from the Technical University of Denmark recently released a new large-scale model product that can predict the probability of human death

The model, called Life2vec, was trained on a unique and rich dataset, said team leader Sune Lehmann Jørgensen. This data set covers the education, medical treatment, diagnosis, income and occupation of 6 million people from 2008 to 2020

The research team said they conducted an experiment and collected data from a group of people aged between 35 and 65 years old. Half of the people in this data died between 2016 and 2020. The researchers asked Life2vec to predict which people were alive and who had died. The results show that Life2vec is 11% more accurate than existing AI models or actuarial life tables used in the financial industry to price life insurance policies

Life2vec Not only that, it can predict personality test results for subsets of the population more accurately than specially trained artificial intelligence models

Sune Lehmann Jørgensen believes that the model has been fully trained on the data and can therefore reveal universal health and social themes. This means it could be used to predict health problems and detect them early, or by governments to reduce inequalities - potentially opening up a whole new model of assessment for future public health science

The rewritten content is: Related links: https://www.newscientist.com/article/2408480-ai-trained-on-millions-of-life-stories-can-predict-risk-of-early- death/

END

In 2023, PHP development stagnated

The following content includes the latest open source information, software updates, technical information and other information

Click here ↓↓↓ Remember to follow ✔ Star⭐ Oh~

The above is the detailed content of A new generation of AI large-scale models for predicting the time of human death. For more information, please follow other related articles on the PHP Chinese website!