TIHM dataset

ITGPT model tackles irregular timeseries data with generative pretraining

Researchers have developed ITGPT, a novel attention-based architecture designed to process multimodal and irregularly sampled timeseries data. This model can be trained using both self-supervised learning and generative…