Large language models, multimodal AI, trustworthy and responsible AI, AI for science, efficient training and inference, evaluation