The creator of LIFE2VEC wants to explore the so -called deep learning plan to discover and predict a series of healthy or social “life events” model and relationship, which is definitely not a pathological obsession.
Professor of Technical University of Denmark, Professor Sune Lehmann, is one of the authors recently published in the “Nature Computational Science” journal.He told AFP: “This is a universal framework for predicting human life. As long as there is training data, it is unpredictable.”
For Lyman, there is infinite possibilities.
He said: “You can predict the health status, so you can predict fertility or obesity, or maybe who will suffer from cancer and who will not. You can even predict whether you will make a lot of money.”
This algorithm uses programs similar to ChatGPT, but analyzes variables that affect life such as birth, education, social benefits, or even work arrangements.
This team is trying to transform this innovation so that the algorithm of language processing can “explore the development and predictability of human life according to the detailed event sequence.”
Lyman said: “From a certain perspective, life is just an event sequence: birth, watching pediatricians, starting school, migration, marriage, etc.”. “
The LIFE2VEC model is based on the anonymous data of about 6 million Denmark collected by Statistics Denmark.
By analyzing a series of events, it may predict all stages of life until the last breath is swallowed.
The accuracy of the death of this algorithm predicting cases is 78%, and the accuracy rate of whether someone will move to other cities or the country is 73%.
However, this tool is still immature if it is used in academic research.
Ryman said: “As of now, it is just a research plan. We are exploring what may be and impossible.”
