Class Teaching Notes

Need

• additional students to sign up for the class: Message former professors, make signs, message Melissa Rose
• a place to host the class: need a projector. Better if I’m not on a separate stage. Nice to have multiple camera
• GCP/AWS/Azure/etc. compute resources for the class: Prefer $1000 credit for preemptable TPU’s
• Cluster time: find friends to run jobs on clusters. Understand the general process.
• a temporary class license for the OpenAI API: I would like access to a temporary license for the OpenAI API so that my students can prepare for the future which they will graduate into when large deep learning models are even more common. We will use the OpenAI API to demonstrate few-shot learning and AI ethics.
• robots to test our code on: I would prefer robots that require relatively little effort to deploy code onto (since this class centers around the algorithmic and numerical aspects of artificial intelligence). However, we will be diving into deep reinforcement learning so I would like to have some complex morphology robots such as articulated arms. (Don’t worry: I will make sure the students smooth motor control outputs) Prefer some have cameras.

Curriculum

• Subdivide curriculum into lecture and lab components
• Add section on MLops and AI ethics
• What is Intelligence. These are all good comments
• Colloquium
• Make accredited

Day 0. (Wed. Sept. 1st) Initialization:

• Welcome!
• high-level overview
  • symbolic computation
  • machine learning
  • deep learning
  • reinforcement learning
  • multi-agent systems

Day 1. (Wed. Sept. 8th) Machines that learn:

• fundamental concepts of machine learning
  • optimization and gradient descent
  • probablistic modeling and factor graphs
• ML techniques
  • regression
  • classification
  • clustering
  • regularization
  • ensembles