04 — Deep Learning
Neural network architectures from fully connected perceptrons through convolutional, recurrent, and transformer models to multimodal systems.
Sublayers
01 — MLP and Representation Learning
Multi-layer perceptrons, activation functions, universal approximation, backpropagation.
02 — Convolutional Networks
Convolution, pooling, CNN architectures (LeNet, VGG, ResNet), object detection, face recognition, neural style transfer.
03 — Sequence Models
Recurrent networks, LSTMs, GRUs, vanishing gradients, trigger word detection.
04 — Transformers
Attention mechanism, transformer architecture, word embeddings, sequence-to-sequence, overview of large-scale transformers.
05 — Multimodal Models
Vision-language models, cross-modal alignment, CLIP-style architectures.