Tutorial 3: RL & Meta-Learning (Done)


Relatively short and fewer code examples (2 main sections and coding only features in the first one). This is fine because the second section is a bit heavy on the PFC as Meta-RL system.

Status: Done

  • Needs a terms / prerequisites section for heavy RL terminology
  • Meta RL
  • Episode
  • Bandit task
  • Tutorial objectives wording (fixed)
  • Needed more explicit connection to understanding generalisation (feels a bit misplaced otherwise) - fixed
  • "Import dependencies" cell doesn't render on the book version of the notebook (why?)
  • Explanations sometimes lacking (added detail) or best removed for clarity if not directly relevant to main point (changes applied)
  • Fixed some idiomatic language use (e.g. "learn on the fly" → "learn as the algorithm is running")
  • Added Big Picture (also kept Summary this time)
  • Ended with some open thoughts / questions
  • Added in refs to upcoming RL content to be introduced later in the course to thread the theme throughout the multiple days
  •