Intelligent Short Text Assessment in eMax

Rapidly increasing student numbers and spreading distance learning systems strengthen the urgent need for using Knowledge Assessment Systems (KAS’s). Recent KAS’s have however, the deficiency of not providing intelligent assessment modules for e.g. the evaluation of freely formulated short answers in-cluding a few sentences or partially solved mathematical problems. The eMax KAS, developed at the Intelligent Knowledge Management Innovation Center of IBM Hungary and the John von Neumann Faculty of Informatics at Budapest Tech, aims to provide these capabilities. Our paper gives an introduction to the in-telligent assessment of short texts component of eMax by presenting the approach used for the formal description of the “answer space” defined as well as the methods chosen for the syntactic analysis, semantic analysis and scoring. Finally, we present the standard course of examination using eMax, then summarize test results and our conclusions.