Neural Networks: From Theory To Practice

Catalog description

This course introduces artificial neural networks (ANNs), which are a subset of machine learning and artificial intelligence. Topics include the ANN mathematics and linear algebra used to model input and output data, optimization algorithms for weight training, the supervised learning process, regression and classification problems, and an introduction to deep neural networks, including their applications in computer vision and natural language processing. Assignments include programming ANNs algorithmically in MATLAB and from built-in machine learning libraries in Python. Permission of instructor required. Prerequisite(s): ENGR 15100 FOR LEVEL UG WITH MIN. GRADE OF D- AND ECE 15200 FOR LEVEL UG WITH MIN. GRADE OF D- AND MA 26100 FOR LEVEL UG WITH MIN. GRADE OF D- Course Learning Outcomes 1. Identify the definitions of and differences among artificial intelligence, machine learning, deep learning, data science, artificial neural networks (ANNs), and other related terminology. 2. Develop the general concepts and mathematics behind neural network regression. 3. Analyze and derivation of local optimization via backpropagation to linear regression and ANN problems. 4. Apply ANNs for supervised learning by properly adjusting algorithmic parameters and examining the training and validation error. 5. Implement and compare multiple local and global optimization techniques. 6. Implement and apply ANNs for classification using logistic regression and SoftMax. 7. Program ANNs algorithmically using MATLAB and from pre-packaged tools using Python. 8. Apply deep learning networks, including convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to real-world applications. 9. Solve real-world data science problems using deep learning tools. View Class Schedule