17-644 Applied Deep Learning
6 units, Carnegie Mellon University
Deep neural networks have made inroads in virtually every industry, propelled by exponential increases in compute power and fundamental progress in modeling. Knowledge of these models is fast becoming a key asset for software engineers, as current systems are quickly starting to include many neural components, and the practice of software engineering itself is starting to benefit from neural program assistance incl. automated bug finding, translation between programming languages. This course equips the next generation of software engineers with knowledge of neural models, the software engineering challenges involved in using these, and handson experience with their applications. It teaches both a rich vocabulary of general, essential concepts including architectures, and recent work on applications of these models, aimed primarily at applications for and in software engineering itself. The course includes a group project aimed at constructing a neural solution for an existing application that will be used to teach the various stages and their pitfalls of building and deploying deep learners.
Prerequisites: none listed
Courseweave needs JavaScript to show sections, workloads and planning tools.