A Tool for Automated Generation of Domain Specific Symbolic Models From Texts

Recent studies in the area of pervasive computing have shown that it is possible to create systems which combine symbolic models incorporating domain knowledge and sensor observations in order to perform tasks such as activity recog-nition in health monitoring use-cases. Creating symbolic models manually can, however, be a very time consuming endeavour and in some cases it also requires experts in the domain of interest. In this work, we showcase a tool for automatic generation of symbolic models from texts written in natural language. The output of our tool is a model description in Planning Domain Definition Language (PDDL). We concentrate on instructional texts in the cooking domain (cooking recipes), but our tool can be also applied in other areas of interest.