The University of British Columbia

3 Credits

EECE 571T

Course Description

• This course provides an introduction to essentials of machine learning, deep learning algorithms and tools, and their applications for engineers and engineering problems.
• This course has no graduate prerequisites or co-requisites. Knowledge of probability, statistics, and programming is required for this course.

Course Topics

1. Clustering and dimensionality reduction algorithms

• K-nearest neighbours algorithm
• Principal Component Analysis (PCA)
• Partial Least Squares (PLS)
• Discriminant Analysis
• Isometric Mapping (ISOMAP)
• Local Linear Embedding (LLE)
• Multidimensional Scaling (MDS)

2. Classification and regression algorithms

• K-means algorithm
• Support vector machines
• Naive Bayes classifier and decision trees
• Hierarchial clustering
• Random forests
• Linear and nonlinear least squares
• Logistic regression
• Kernels and regularization

3. Introduction to deep learning

• Neurons, neural networks, multi-layer perceptron (MLP)
• Convex optimization, stochastic gradient decsent (SGD)
• Backpropogation
• Transfer learning, augmentation, regularization
• Introduction to Tensorflow, Keras, Google Colab

4. Deep learning in images and time series

• Convolutional neural networks (CNNs)
• Applications of CNNs: image classification, object/abnormality detection, segmentation, attention models, etc.
• Recurrent Neural Networks (RNNs) Long-short Term Memory (LSTM), Gated Recurrent Unit (GRU)
• Applications of RNNs: forecasting, time series and video analysis, natural language processing (NLP)

5. Advanced machine learning algorithms

• Generative Adversarial Networks (GANs) _ Variational Auto-Encoders (VAEs)
• Active learning, online learning _ Self-supervised learning, semi-supervised learning,
• Label propagation, weak labelling
• Reinforcement learning

Course Assessment

• Assignments: 40%; four assignments in total.
• Course project: 60%; project proposal, oral presentation, and written report.