Advanced Deep Learning

A graduate level deep learning course focusing on attention and transformer architectures. Introduces selection of transformer-based models applied to language, vision and audio problems, generative, diffusion, state-space, and mixture of expert models. Analysis of zero-shot learning, hallucinations, graph neural networks, multimodal architectures, and agents. Multiple object detection methods are compared. Course is executed in hybrid presentation/flipped-classroom styles, with practical exercises and student presentations on design and operation of DL models.