Cross-Modal Information Retrieval Tasks | State-of-the-Art