XMolCap: Advancing Molecular Captioning through Multimodal Fusion and Explainable Graph Neural Networks | IEEE Journals & Magazine | IEEE Xplore