Welcome to the Multimodal Analysis Group (MagCIL) of the Institute of Informatics and Telecommunications, National Center for Scientific Research "Demokritos".

We conduct research in multimodal deep learning, with a focus on audio, speech, music, and visual analysis and generation. Our work spans self-supervised and few-shot representation learning, generative modeling, adversarial robustness, computer vision, remote sensing, and multimodal fusion, from fundamental research to deployed AI systems in health, culture, defence, and environmental monitoring.