Detección automática de errores en el Corpus SenSem

The goal of the SenSem project is the syntactico-semantic annotation at verbal level of a big corpus of Spanish. One of the phases in this process consists in correcting annotation errors in the corpus. To that aim, we are developing some procedures to detect errors automatically and correct them. These procedures are currently being studied and designed, before they are actually implemented computationally. In this paper we describe the project and the kind of annotation it involves, and we discuss a typology of errors, as well as some detection procedures.