“…The problem's text is sent to the NLU agent where the available information is extracted through words segmentation, semantic analysis and discourse clustering (Li, Zhao, Li, & Cheng, 2013;Li, Zhao, Yi, Cheng, 2011;Xu, Zhao, Yi, Li, Cheng, 2012;Xu, Zhao, Cheng, 2010). Afterwards, the agent selects the known conditions and unknown conditions from the information.…”