3 credits
Spring 2025 Lecture Upper Division
Teaches the foundation of neural network (aka Deep Learning) approaches and applications to medical imaging: or how to create computable models of biological neural systems, in particular large-scale neural networks, for processing medical images and data. Drawing inspiration from neuroscience and statistics, Deep Learning introduces the basic background on neural networks, presents computable neuron models and extends to large networks of neurons. Students study deep learning with both real and artificial neural networks, along with methods of deep learning about the environment by applying back propagation, Boltzmann machines, auto-encoders, convolutional neural networks and recurrent neural networks. Students will explore how Deep Learning is impacting our understanding of intelligence and contributing to the practical design of automatic medical imaging tools that are able to augment expert medical personnel, enhance diagnosis accuracy and speed, and lower health-care costs.
Learning Outcomes1Understand and simulate a large biological neural network with several million of neurons.
2Design and use a bio-inspired deep-learning network for application in medical imaging.
3Write algorithms that learn to segment, track, categorize, and classify objects of interest.
Prerequisites
Attachments
made by Purdue students.