Vision-Guided Chunking Is All You Need: Enhancing RAG with Multimodal Document Understanding - Explained Simply | ArXiv Explained