Advanced Machine Learning

National Research University Higher School of Economics

N.B.: Please don't use the assignment and quiz solution at first time, only use when you get stuck really bad situation. Try to solve the problem by yourself.

About this Specialization

This specialization gives an introduction to deep learning, reinforcement learning, natural language understanding, computer vision and Bayesian methods. Top Kaggle machine learning practitioners and CERN scientists will share their experience of solving real-world problems and help you to fill the gaps between theory and practice. Upon completion of 7 courses you will be able to apply modern machine learning methods in enterprise and understand the caveats of real-world data and settings. - Source

Assignment - Offline Solution Instruction

• Your own hardware: https://github.com/hse-aml/intro-to-dl#offline-instructions
• Google Colab (cloud): https://github.com/hse-aml/intro-to-dl#running-on-google-colab-tested-for-all-weeks

Resources

• More about matrix derivatives, here's what you can read:
  1. https://compsci697l.github.io/docs/vecDerivs.pdf
  2. https://www.math.uwaterloo.ca/~hwolkowi/matrixcookbook.pdf
• Keras reading: https://blog.keras.io/keras-as-a-simplified-interface-to-tensorflow-tutorial.html

Course - 1

Introduction to Deep Learning - Coursera - GitHub - Certificate

Table of Contents

• Week 1
  • Lesson Topic: Linear regression and classification, Gradient descent, Linear models, Overfiting, Validation, Regularization, Stochastic gradient descent, Optimization
  • Quiz: Linear models, Overfiting and Regularization
  • Assignment: Linear models and optimization
• Week 2
  • Lesson Topic: MLP, Chain rule, Backpropagation, Matrix derivatives, TensorFlow framework, Keras,
  • Quiz: Multilayer perceptron, Matrix derivatives
  • Assignment: MNIST digits classification with TF
  • Optional: Your very own neural network
• Week 3
  • Lesson Topic: Convolutional layers, CNN architecture, Computer Vision tasks
  • Quiz: Convolutions and pooling
  • Assignment: Your first CNN on CIFAR-10, Fine-tuning InceptionV3 for flowers classification
• Week 4
  • Lesson Topic: Unsupervised learning, Autoencoders, NLP, Word embeddings
  • Quiz: Word embeddings
  • Assignment: Simple autoencoder
  • Optional: Generative Adversarial Networks
• Week 5
  • Lesson Topic: Recurrent layers, Simple RNN and Backpropagation, LSTM, GRU, Practical use cases for RNNs
  • Quiz: RNN and Backpropagation, Modern RNNs, How to use RNNs
  • Assignment: Generating names with RNNs
• Week 6
  • Lesson Topic: None
  • Quiz: None
  • Assignment: Image Captioning Final Project

Course - 2

How to Win a Data Science Competition: Learn from Top Kagglers - Coursera - GitHub - Certificate

Table of Contents

• Week 1
  • Lesson Topic:
  • Quiz:
  • Assignment:
• Week 2
  • Lesson Topic:
  • Quiz:
  • Assignment:
• Week 3
  • Lesson Topic:
  • Quiz:
  • Assignment:
• Week 4
  • Lesson Topic:
  • Quiz:
  • Assignment:
• Week 5
  • Lesson Topic:
  • Quiz:
  • Assignment: