There are several ways that vocabulary differences are compensated for by Discovery engines. First, like traditional Search engines, Discovery services can accommodate synonym dictionaries, which essentially perform searches using the input term/phrase and its equivlents. That assumes the synonyms have been defined and incorporated into the system.
For the most part, however, Discovery services can usually generate good results even without these synonyms. Because modern Discovery services are based on full text indexes, they use more than metadata in the calculation of relevance. While the metadata may use specific terms to describe the content, it's highly likely that the full-text will use a wider variety of terms. If it does, the presence of those terms will make the article/holding more relevant and rank it more highly in the search results.