Advanced Machine Learning

Chennai Mathematical Institute - Autumn 2022

• Instructor: Pranabendu Misra
• Teaching Assistants: Somenath Dake and Ameya Kamat
• Time: Tuesday and Friday 11:50AM
• Venue: Lecture Hall 803


• References:

  • Avrim Blum, John Hopcroft and Ravi Kannan: Foundations of Data Science, Cambridge University Press 2021 Available Here

  • Ian Goodfellow, Yoshua Bengio and Aaron Courville: Deep Learning, MIT Press 2016 Available Here

  • Aston Zhang, Zack C. Lipton, Mu Li, Alex J. Smola: Dive into Deep Learning, 2022 Available Here

  • Richard S. Sutton and Andrew G. Barto: Reinforcement Learning: An Introduction, MIT Press (2nd ed) 2018 Available Here

  • Aurélien Géron: Hands-On Machine Learning with Scikit-Learn, Keras and Tensorflow (2nd ed), O'Reilly 2019

  • Francois Chollet: Deep Learning with Python, Manning Publications 2017

  • Nikhil Buduma: Fundamentals of Deep Learning, O'Reilly 2017

  • Stuart J Russell and Peter Norvig: Artificial Intelligence: A Modern Approach, Pearson (3rd ed) 2016

  • Daphne Koller and Nir Friedman: Probabilistic Graphical Models – Principles and Techniques, MIT Press 2009

Announcements

• Please subscribe to the course on moodle

Lectures

• Lecture 1
  Theoretical foundations of ML (pdf),

  PAC learning, uniform convergence

  Representational capacity, overfitting, underfitting, hyperparameter search

  • Blum et al, Chapter 5.1, 5.4, 5.4.1
  • Goodfellow et al, Chapter 5.2



• Lecture 2
  (VC-Dimension (pdf), Loss Functions (pdf),

  VC dimension

  • Blum, Hopcroft and Kannan, Chapter 5.5, upto 5.5.2

  Choosing loss functions: MLE, cross entropy

  • Goodfellow et al, Chapter 5.5, 3.13



• Lecture 3 and 4: Neural Networks and Training
  Class Notes (pdf),

  Neural networks, unstable gradients, initialization strategies, non-saturating activation functions, batch normalization

  • Géron, Chapter 11
  • Chris Olah: Backpropogation
  • Boza Barak: Yet another backpropagation tutorial

• Lecture 5: Training Neural Networks II
  (Class Notes (pdf),

  ill-conditioning; optimizing backpropagation: momentum, adaptive learning rates; regularization; local minima; model identifiability; saddle points

  • Buduma, Chapter 4
  • Géron, Chapter 11



• Lecture 6: Fully Connected Neural Networks for MNIST
  (DNN for MNIST, ipynb,
  DNN for CIFAR 10, ipynb)




• Lecture 7: Convolutional Neural Networks
  (Class Notes(pdf),

  • Géron, Chapter 14
  • Buduma, Chapter 5
  • Goodfellow et al, Chapter 9, 9.1, 9.2, 9.3, 9.5



• Lecture 8: CNN for MNIST, CIFAR-10 and Adversarial attacks
  (CNN for MNIST, ipynb,
  CNN for CIFAR 10, ipynb
  Fast Gradient Sign Method attack for CNN for CIFAR 10, ipynb)

  • Dive into Deep Learning, Chapter 6
  • Deep Learning: ConvNet Evolutions, Architectures, Implementation Details and Advantages
  • Fast Gradient Sign Attack



• Lecture 9: Recurrent Neural Networks
  (Slides(pdf),

  • Géron: Chapter 15
  • A Critical Review of Recurrent Neural Networks for Sequence Learning, Zachary C. Lipton, John Berkowitz, Charles Elkan
  • Understanding LSTM Networks



• Lecture 10: Examples of RNN and LSTM
  (RNN and LSTM Example, ipynb)




• Lecture 11: Autoencoders and Variational Autoencoders
  (Autoencoders (pdf),

  • Introduction to autoencoders, Alfredo Canziani
  • Generative Models – Variational Autoencoders, Alfredo Canziani
  • Variational Autoencoders are Beautiful, Steven Flores
  • Chapters 14 and 20.10.3, Goodfellow et.al.



• Lecture 12: Examples of Autoencoders and Variational Autoencoders
  (Autoencoders, VAE)




• Lecture 13: Generative Adversarial Networks
  (GAN (pdf), GAN Example)

  • Deep Generative Models, Ava Soleimany (pdf,
  • Chapter 18, Dive into Deep Learning (Link to Chapter)



• Lecture 14: Reinforcement Learning
  (Slides)

  Introduction to reinforcement learning, multi-armed bandits

  • Sutton and Barto, Chapters 1, 2.1–2.7
  • Russell and Norvig, Chapter 17.1

• Lecture 15: Markov Decision Process
  (Slides),

  Markov Decision Processes: Basic definitions and examples, policies and value functions, Bellman equation, optimal policies

  • Sutton and Barto, Chapter 3

• Lecture 16: Dynamic Programming Methods for RL
  (Slides)

  Markov Decision Processes: Policy evaluation, policy iteration, value iteration

  • Sutton and Barto, Chapter 4

• Lecture 17: Monte Carlo Methods for RL
  (Slides)

  Markov Decision Processes: Monte Carlo methods

  • Sutton and Barto, Chapter 5

• Lecture 18: Temporal Difference Learning
  (Sildes)

  Markov Decision Processes: Temporal Difference Learning

  • Sutton and Barto, Chapter 6

• Lecture 19: Deep Reinforcement Learning
  (Slides, Deep Q Learning (Example).html,

  Deep reinforcement learning

  • Reinforcement Learning, lecture by Fei-Fei Li, Justin Johnson and Serena Yeung (Slides, Video)
  • Playing Atari with Deep Reinforcement Learning, Mnih et al

• Lecture 20: Bayesian Optimization
  (Slides)

  Bayesian Optimization, Gaussian Process Regression

  • Example of Bayesian Optimization (ipynb) Sourish Das
  • Tutorial on Gaussian Process Regression, Richard Turner
  • More on Multivariate Gaussians
  • A Tutorial on Bayesian Optimization, Peter I. Frazier
  • A Tutorial on Bayesian Optimization, Brochu et al
  • Gaussian Processes for Regression: A Quick Introduction, M. Ebden
  • Bayesian optimization with skopt