In a groundbreaking advancement, researchers at Danmarks Tekniske Universitet (DTU) have developed a novel algorithm, named life2vec, capable of predicting the time of death with an impressive 78% accuracy.
This technology, drawing inspiration from the language model ChatGPT, marks a significant stride in predictive analytics and could transform the healthcare industry.
Professor Sune Lehmann Jørgensen, a leading figure at DTU, explains that life2vec aims to enhance our understanding of the events in human life by analyzing a variety of data. Its potential application in healthcare is promising, especially for early detection and intervention in diseases, potentially saving numerous lives.
Described by the media as an 'AI death calculator', life2vec employs techniques similar to those used in models like ChatGPT. However, instead of processing textual data, it analyzes life events, ranging from birth and education to professional life.
The research team focused initially on mortality due to the abundance of insurance data available in this area. They examined data from 2008 to 2016 to ascertain if the algorithm could predict the number of participants alive in 2020. The results showed a remarkable 78% accuracy in its predictions.
Prof. Jørgensen emphasizes that this development is just the beginning of a long journey aimed at predicting various aspects of our society.
The algorithm was fed with detailed life event information, assigning digital values to each for analytical processing. For example, it could receive data like 'Francisco earned 20,000 kroner monthly in 2012 as a guard at Kronborg Castle in Helsingør' or 'Hermione had five different A-level subjects in high school.' These intricate details significantly influence life2vec's predictions.
The algorithm's calculations consider the sequence and minutiae of life events, which are crucial for accurate forecasting. This approach could revolutionize predictive modeling in various fields, with a particular impact on healthcare.
DTU's life2vec stands as a testament to the potential of AI in enhancing our understanding of human life and its various trajectories, opening up new avenues for predictive analytics in healthcare and beyond.
