Researchers from Denmark and the US have developed a doom calculator called "Life2vec" that predicts when a person dies.
How Life2Vec Doom Calculator Works
The model grounded in artificial intelligence is somewhat similar to ChatGPT, working as a chatbot. However, the interactions made with Life2vec are quite different compared to those made with ChatGPT.
Life2vec covered data of over 6 million individuals across Denmark. The data was granted by the government of the country, as it collaborated on the study. Such data includes education, age, jobs, income, health, and other events.
Interestingly, each data piece was assigned with varied digital tokens. These were all categorized specifically. For example, the representation for a forearm fracture was "S52," while having a job at a tobacco shop had the code "IND4726."
The AI model works by assimilating life data about individuals in sentences. As the model evolved, it gained the capability to build trajectories of individual human lives.
Sune Lehmann, an author of the study and a professor of networks and complexity science from the Technical University of Denmark, explains that the entire story of human life could be perceived as a huge long sentence covering different things that may happen to an individual.
Interestingly, the algorithm was also capable of predicting the personalities of individuals. To accomplish such a feat, the team trained the model to predict the answers of people on a specific personality test. However, Lehmann warned that since the data were from Denmark alone, the predictions could not be applicable to those who live in other areas.
ALSO READ : Doom Calculator 'Life2vec' Algorithm Can Predict Person's Earnings, Time of Death With 78% Accuracy
How Accurate Is Life2vec?
When it came to "time of death" predictions, the AI was eventually able to correctly predict those who died in 2020 with an accuracy of 78%.
The study participants were not informed of their death predictions, as this would be something "very irresponsible," Lehmann notes. The professor adds that the team is hoping to share result details in a way that honors the privacy of the participants.
Lehmann adds that Life2vec can shed light on the factors that could help extend life span. Though this is something the researchers have not gone deeper into, it is another crucial application of the model.
More specifically, the study found that some factors that are linked to a longer life are having a leadership role or earning more. On the other hand, factors that could contribute to a shorter life include a mental health diagnosis, being male, and having a skilled profession.
Check out more news and information on Medicine & Health in Science Times.
