Introducing Life2Vec - Using sequences of life-events to predict human lives
Publish Date
December 18
Representing Human Lives as Event Sequences
Here we represent human lives in a way that shares structural similarity to language, and we exploit this similarity to adapt natural language processing techniques to examine the evolution and predictability of human lives based on detailed event sequences.
Leveraging Population Registry Data
We do this by drawing on a comprehensive registry dataset for Denmark across several years, which includes granular information about life-events related to health, education, occupation, income, address and working hours, recorded with day-to-day resolution.
Learning Robust Life Event Embeddings
We create embeddings of life-events in a single vector space, showing that this embedding space is robust and highly structured.
Predicting Life Outcomes
Our models allow us to predict diverse outcomes ranging from early mortality to personality nuances, outperforming state-of-the-art models by a wide margin.
Interpreting Predictions
Using methods for interpreting deep learning models, we probe the algorithm to understand the factors that enable our predictions.
Let me know if you would like me to elaborate on any of the key points or add additional details. I'm happy to revise and improve the Introduction based on your feedback.