Part II: Deep Learning Architectures

Modern deep learning: convolutional networks, transformers, autoencoders, and generative models for scientific applications.

CNNs for Image & Signal Data

Convolution, pooling, ResNets, transfer learning

RNNs & Transformers

LSTM, attention mechanism, self-attention

Autoencoders

Dimensionality reduction, denoising, variational

Generative Models (VAEs & GANs)

Latent spaces, adversarial training, diffusion models

← Part IPart III →