About the Course

Deep Learning uses a cascade of multiple layers of nonlinear processing units for feature extraction and transformation. Each successive layer uses the output from the previous layer as input. The Deep Learning (Level 2) has the following contents.

Main Contents

Recurrent Neural Networks 
  • RNN
  • LSTM
  • GRU 
  • Time series
  • Language modeling 
  • Image captioning, visual question answering
Attention Mechanism
  • Encoder and Decoder
  • Attention idea 
  • Attention score functions
  • Self-attention
  • The transformer
Generative Adversarial Networks
  • GAN framework
  • GAN training
  • GAN applications
Unsupervised Learning
  • Likelihood-based Models
  • Latent Variable Models
  • Unsupervised Learning applications