Principal investigator (ÚFAL): 
Grant id: 


Information mining from spoken dialogue

The quality of dialogue systems depends on the amount of knowledge which can be provided to its users. Large knowledge bases (DBPedia, Freebase) still suffer from data sparsity because the information contained in the databases become obsolete as new facts appears. On the other hand, specialised, highly updated databases restrict the dialogue systems to providing the knowledge without broader context. The goal of the project is to develop strategies for extracting new knowledge from users, and thus learn either completely new pieces of information or enrich known facts.