Deep Learning Networks

This course explores MLP (NN1.0), CNN (NN2.0) and Transformers. Basic topics: Sigmoid/ReLU activations, BP learning, dropout, regularization, generalization, classification and prediction. Advanced topics: (1) unifying MLP and CNN learning methods, (2) unifying classification and regression applications, and (3) input/output residual learning to mitigate curse of depth, (4) Hybrid NAS (Progressive & Regressive Neural Architecture Search) and (5) Generative AI, via transformer and stable diffusion, which can learn contextually from huge pretraining datasets by using Large Language Models (LLM), enabling generation of creative contents.