Multimodal and Cross-Modal Learning Techniques

Multimodal data fusion is increasingly becoming relevant in machine learning, enabling richer, more comprehensive, and robust models. This is achieved by effectively integrating diverse data types including images, text, audio, sensor signals, and scientific datasets (e.g., crystallographic, spectroscopic, and electronic properties measurements). This special issue invites original research articles and reviews focusing on cutting-edge advancements in multimodal alignments and cross-modal learning techniques. Submissions are encouraged to explore novel methodologies, theoretical foundations, and practical implementations addressing key challenges in multimodal integration, such as handling modality heterogeneity, scalability, interpretability, and representation learning across diverse data sources.