Skip to content

Latest commit

 

History

History
33 lines (25 loc) · 2.57 KB

24784.md

File metadata and controls

33 lines (25 loc) · 2.57 KB
Raw

24-784: Trustworthy Intelligent Autonomy

Category Difficulty
Projects 3/5
Paper Reviews 4/5

This course has five modules, consisting of adversarial robustness, reinforcement learning, AI safety, digital twin/metaverse/certification, generalization, and social good (privacy and fairness). You get to understand the basics in these areas, recent progress, and limitations of existing methods. Also you will explore the novelty and potential extension of various state-of-the-art trustworthy AI research. Knowledge and research skills learned in this course can be applied to self-driving, healthcare devices, assistant robots, and intelligent manufacturing. Particularly, the projects and challenges will mainly be in the field of self-driving.

Expected prior knowledge

Linear algebra, Intro level of Statistics, Intro level of Machine Learning

What to expect

  • Projects:(30%): These are individual and are meant to help you with public datasets, open-source packages, and basic programming for trustworthy intelligent autonomy related problems.
  • Challenges:(30%): You get to work in teams for these and solve two sets of challenging and realistic tasks.
  • Paper reviews:(30%): Throughout the course you will read 15 papers in five popular directions of trustworthy AI. You will be required to write one review for each paper. The review will be either a short version (<300 words) or a long version. You will need to write at least three long reviews. In addition to this, you will select one paper from the long reviews and record a 10-min video of the paper presentation.
  • Exams: There are no exams in this course.

How to do well

  • Make sure you attend and understand lectures. The lectures are not recorded.
  • Start Projects and challenges early.
  • Attend Office Hours (OH), utilize the Support Camps conducted by the TAs for the Projects and Challenges. They are really helpful.
  • Would recommend doing some introductory course on Reinforcement Learning if you haven't already. I found the slides of Berkeley CS 285: Deep Reinforcement Learning, Decision Making, and Control pretty helpful.
  • Follow and participate in discussions on Campuswire.

Resources

  • There is no textbook for this course. Lecture slides are really good and get posted on Canvas.
  • But some good resources to check out include: Stanford CS231n: Convolutional Neural Networks for Visual Recognition Berkeley CS 285: Deep Reinforcement Learning, Decision Making, and Control UIUC CS 598: Adversarial Machine Learning