IS 6422 Deep Learning Architectures 3.0 Credits
The course introduces topics in deep learning that guide how neural networks are designed to meet the representation and prediction needs of different data types, learning tasks, and applications. Presents the development of state-of-the-art architectures and techniques developed for their pre-training and fine-tuning, including for voice transcription, image processing, sequence processing, and graph representation. Covers best practices for building networks with a variety of linear, convolutional, recurrent, and self-attention, and preference-tuning components in Python. Through lectures, programming assignments, and leadership of a team project, students will implement a deep learning architecture for a relevant data science domain.
Repeat Status: Not repeatable for credit
Prerequisites: (IS 5330 [Min Grade: D] or CS 5520 [Min Grade: D] or DSCI 631 [Min Grade: D] or CS 613 [Min Grade: D]) and (IS 5321 [Min Grade: D] or CS 5110 [Min Grade: D] or DSCI 521 [Min Grade: D] or CS 521 [Min Grade: D])