Which challenges are addressed by Caption Production but not Semantic Search?

GQ-0279