Neural networks lie at the heart of many modern technologies, including facial recognition, medical diagnoses, self-driving cars and ChatGPT. This course explores the structure, training, and application of neural networks, with a focus on both building coding proficiency and generating a strong conceptual understanding.
Specific topics will include embeddings, convolutional and recursive networks, transformers, autoencoders, natural language processing, and generative adversarial networks. Students will have the opportunity to work on a project that allows them to train a neural network model for a topic of personal interest, either from another course (e.g. bioinspired robotics or the farm data project) or from some broader set of engagements. The class will emphasize hands-on collaborative learning, fusing theory, coding, writing, and discussion. Students will be evaluated on their participation, weekly exercises, Python coding, and final project.
Since neural networks lie at the interface of computer science and mathematics, a prerequisite for this course is either Computer Science Foundations or Calculus. If you are interested in taking the course but have some other background, contact the instructor directly and ask about whether it would be a good fit.
Anticipated Credit Equivalencies:
*4 - Neural Networks
Registration
Students are expected to have coursework in discrete math, computer architecture, data structures, and one year of computer programming. These prerequisites are covered by completion of Computer Science Foundations and Data Structures and Algorithms, or equivalent courses elsewhere.
Academic Details
Computer Science, including software development, Web development, data science, and IT.
All 4 credits of the work in this program are designed to be upper-division math/science.