Deep Learning for Predictive Analytics. 3 credits

Catalog description

Overview of the theoretical and algorithmic foundations of deep learning as well as practical aspects of developing deep learning predictive models. Topics include theoretical results from convex optimization and approximation theories, first and second order optimization algorithms (stochastic gradient descent, Nesterov acceleration, Newton’s), overview of popular architectures (recurrent and convolutional networks), accelerated linear algebra, GPU computing, automated differentiation and Bayesian inference. Applications in engineering, finance and artificial intelligence. Practical aspects of building predictive models, such as architecture selection, and data normalization. Extensive use of computational tools, such as the Python language, both for illustration in class and in homework problems. In addition to traditional instruction, a number of case studies and students’ model building projects provide further breadth and exposure. Offered by Systems Engr & Operations Rsch . May not be repeated for credit.