Units: 4
Hours: Lecture, 3 hours; discussion, 1 hour
Catalog page 480
Prerequisites: MATH 161A with a grade of C- or better
Description: Introduction to core mathematical principles in deep learning, emphasizing training neural networks for computer vision. Topics cover various architectures, image transformation techniques (segmentation, super-resolution, denoising), object detection, and transfer learning. Also explores key optimization tools and neural network visualization to ground theoretical innovations in deep learning, providing practical insights.