Composed Video Retrieval via Enriched Context and Discriminative Embeddings | Read Paper on Bytez