What process must address acoustic ambiguity in NLP?
Part-of-speech tagging
Word sense disambiguation
Speech recognition
Discourse