Posts of Peril

New model detects hazard information in text, applied to geopolitical events

Researchers have developed a new model to detect information about hazards in text, which is often overlooked compared to sentiment or emotion analysis. This model, trained on a dataset of X posts, performs well and ext…