The symposium aims at bringing together researchers from different strands of semantic processing research, with a focus on two aspects. The first is the textual inference paradigm, proposed as a unifying generic framework capturing major semantic processing needs in different application areas. Textual entailment research has yielded so far a range of inference algorithms, which were mostly developed within individual “in-house” systems. The symposium aims at addressing current bottlenecks by discussing future research directions and perspectives for consolidation and unification of textual inference technology.
The second aspect concerns linguistic structures, as identified by empirical work within corpus linguistics, whose benefit for Computational Linguistics has not yet been fully realized. We are interested in corpus investigation methods which identify the properties of natural languages that can be efficiently computed and used to improve the actual state of the art on NLP tasks involving meaning. The workshop aims at contributing to the common ground needed to create language models that allow the computational linguistic community to take the next step forward in developing natural language applications that process semantic knowledge successfully.