Young Kyun Jang, Donghyun Kim. "Visual delta generation with large multi-modal models enhances composed image retrieval using unlabeled data." Scientific Reports (2025). https://doi.org/10.1038/s41598-025-07798-6