<aside> 💡 Welcome! Below are all the topics and readings we’ll be covering over the 8-week program. For each session, you’ll do ~1 hour of reading and ~1 hour of discussion with an in-person group **(though you’re welcome to read beforehand). No work is expected outside of the weekly 2-hour meeting. We’ve also included additional readings for some topics if you’d like to learn more about them on your own time.
This program is intended to be your first step in exploring a career in technical AI safety research. You’ll get an overview of all the most important concepts related to AI alignment, the catastrophic risks posed by advanced misaligned AI, and promising research agendas to mitigate these risks.
Apply here by Saturday, September 14
</aside>
We’ve included various resources below to help you brush up on key concepts in Machine Learning. We will not be covering these in-person so you are expected to look at these on your own time if needed. If you’ve taken CSCI14xx or have significant past experience with ML/DL, you can probably safely skip this.
If you feel concerned about being prepared, please reach out to one of the organizers ([email protected] or [email protected]).
A short introduction to Machine Learning (Richard Ngo, 2021)
3Blue1Brown Neural Networks (all except Chapter 7 are useful)
Machine Learning for Humans, Part 2.1 (Vishi Maini, 2017)
What is self-supervised learning? (CodeBasics, 2021)
Introduction to reinforcement learning (Hado van Hasselt, 2021)
Transformers, explained: Understand the model behind GPT, BERT, and T5 (Google, 2021)
What are the current trends in AI? What is the role of AI safety and why is it important?