Teaching

Mathematics for Machine Learning (Autumn 2021 @ Imperial)

  • Gradient descent (GD) [notes]
    • constant step-size GD, convergence analysis [slides]
    • Advanced techniques & SGD [slides]
  • Ridge regression [notes]
    • bias-variance trade-off [slides]
  • Principal component analysis (PCA) [notes]
    • minimum reconstruction error view [slides]
    • maximum variance view [slides]
  • Constrained optimisation
  • The kernel trick [slides][notes]

Deep Learning (Winter 2022 @ Imperial)

  • Generative models
    • Intro to generative modelling [slides]
    • Variational auto-encoders [slides] [notes]
    • Generative adversarial networks [slides] [notes]
    • Recent advances in deep generative models [slides]
  • RNNs
  • Attention networks
    • Attention basics & Transformer [slides] [notes]
    • Recent advances in attention methods [slides]