STAT 487
Download as PDF
Deep Learning
Statistics
College of Computational, Mathematical, & Physical Sciences
Course Description
Examination of deep learning models with an emphasis on statistical principles such as model selection, model interpretability, tuning, and generating prediction uncertainty. Deep learning methods, explainable AI, backpropagation, transfer learning, and training methods. Analyses on real-world applications such as natural language, audio, and image processing.
When Taught
Winter
Min
3
Fixed/Max
3
Fixed
3
Fixed
0
Title
Deep learning methods
Learning Outcome
Students will be able to select the appropriate deep learning method to use for various types of data and learning (e.g., supervised vs. unsupervised) needs.
Title
Software implementation of deep learning
Learning Outcome
Students will be able to implement using standard software the following deep neural networks, using simple neural networks as well making use of more complex pre-trained neural networks: convolution neural networks (CNN), recurrent neural networks (RNN), and transformers.
Title
AI methods
Learning Outcome
Students will be able to use explainable AI methods to provide intuition about model behavior and to obtain insights about the data.
Title
Natural language processing
Learning Outcome
Students will understand the basics of natural language processing and how to effectively use deep learning models to parse, clean, understand, and predict language analyses.
Title
Signal processing
Learning Outcome
Students will be able to apply methods to various signal processing scenarios such as image processing, audio processing, and/or language processing.
Title
Fundamental of deep learning
Learning Outcome
Students will understand enough of the fundamental concepts and workings of various deep learning methods to make informed choices about model and parameter settings for specific applications.
Title
Uncertainty quantification for deep learning
Learning Outcome
Students will be able to understand the current state of uncertainty quantification techniques and research in deep learning.
