Course Description

Enabled by large datasets, new algorithms, and hardware improvements, Artificial Intelligence is taking on increasingly complex tasks, in some cases transforming how scientific research is done. Building on the Machine Learning foundations from BIOF 509, BIOF 510 will explore how these recent advancements have enabled predictive algorithms to perform tasks that once required human intelligence. We will examine how different architectures and training regimens can tackle different types of data and problems, from basic multi-layer perceptrons to convolutional and recurrent neural networks, autoencoders, and Generative Adversarial Networks. In addition to looking at successful applications in the biomedical domain, we will also discuss the hazards of deploying Artificial Intelligence in science and best practices for ensuring trustworthy results. Concepts will be reinforced by discussions, quizzes, coding assignments, and a research project, in which students can apply the methods from the course to a dataset of their choice. Coding will be in Python and will use popular Deep Learning frameworks.

Learner Outcomes

When you complete the course successfully, you will be able to:

• Explain how neural networks are able to identify patterns in data
• Choose an appropriate Deep Learning architecture for a given problem
• Implement common Deep Learning architectures in Python using popular libraries
• Follow best practices for acquiring and preparing data for Deep Learning
• Interpret the output of Deep Learning systems

Microcredential(s)

This course applies toward the Bioinformatics Endeavor digital badge.

Textbook Information

A textbook is required or recommended for this course.

Click here to view a textbook list for FAES courses and purchasing information. Please note that tuition does not include textbooks.