CSU44061/CS7CS4: Introdiction to Machine Learning 2022-23

Module descriptor

Lectures: Tues 5-6pm, Fri 11-12noon
Lecturer: Doug Leith and Giovanni Di Liberto
Teaching Assistants: Dilina Rakapakse, Mayank Soni, Chalehchaleh Amirhossein

The module is a first course in machine learning. Its quite hands-on and aims toget you to the point where you can apply common ML techniques in a solid way and critically evaluate performance.

• Introduction, incl movie sentiment analysis example
  • video- movie example
• Linear Regression & Gradient Descent
  • video - linear regression 1
  • video - linear regression 2
• Logistic Regression & SVMs
  • video - logistic regression 1
  • video - logistic regression 2
  • video - support vector machines
• Feature Engineering
  • video - intro to feature engineering
  • video - under/over-fitting
• Cross-validation & Regularisation
  • video - cross-validation
  • video - hyperparameter tuning
  • video - model selection
• Evaluating Performance
  • video - using a baseline
  • video - classification performance
  • video - roc curves and auc
  • video - model assessment
• k-Nearest Neighbours, Decision Trees
  • video - k nearest neighbours 1
  • video - k nearest neighbours 2
• Neural Nets
  • video - neural nets 1
  • video - stochastic gradient descent 1
  • video - stocastic gradient descent 2
  • video - neural nets 2
• Kernel Trick
  • video - kernel trick 1
  • video - kernel trick 2
• Text Features
  • video - text features 1
  • video - text features 2
  • video - text features 3 TFIDF
  • video - text features 4
• Time-series Features
  • video - time series 1
  • video - time series 2
  • video - time series 3
• Convolutions
  • video - what is convolution
  • video - edge detecttion
  • video - more on convnets
• Convolution Net Layers
  • video - convolutional layer
  • video - pooling and FC layers
  • video - convnet example
• Convolution Net Architectures
  • video - why convolutions?
  • video - LeNet and AlexNet
  • video - VGG16/19 and ResNet
  • video - transfer learning
• Convolution Net Applications
• Recommender Systems - Item-based
  • video - recommenders: simplified setup
  • video - content and item-based recomenders
  • video - session-based recommenders
• Recommender Systems - Matrix Factoprisation
• Unsupervised Learning k-Means
  • video - unsupervised vs superrvised learning
  • video - k-means 1
  • video - k-means 2
  • video - k-means news example

• Coursera Machine Learning Module
• The Elements of Statistical Learning
• Statistical Methods for Computer Science Module