The Multimodal Data-Driven Optimization and Learning (MultiDOL) workshop focuses on addressing the growing need for integrating diverse data modalities—such as images, text, and sensor data—into evolutionary learning and optimisation frameworks. As real-world problems increasingly involve heterogeneous data sources, from autonomous systems combining visual and LiDAR data to healthcare applications fusing medical images and clinical records, traditional learning and optimization approaches must evolve.

This workshop explores novel methods for multimodal data integration in evolutionary algorithms, including multimodal representation learning, cross-modal feature fusion, and hybrid EC-deep learning approaches. We welcome contributions on theoretical foundations, algorithm design, benchmark development, and applications across areas, such as robotics, smart manufacturing, healthcare, and environmental monitoring. We brings together researchers from evolutionary computation, machine learning, computer vision, and application domains to address key questions: How can evolutionary algorithms effectively process information from multiple data modalities? What new optimization paradigms emerge when combining EC with modern multimodal AI including foundation models? How do we design appropriate benchmarks for these problems?

MultiDOL aims to foster collaborations between the EC community and multimodal AI researchers, establishing a foundation for sustained research in this emerging interdisciplinary area.