We present an early version of a method for open-domain semantic assertion extraction from natural language texts. To combat the shortage of training data for the task, a two-stage pipeline is employe
