Interplay Between Reinforcement Learning and Generative Artificial Intelligence (Fall 2025)

Four problem sets, each worth 10% of the course grade.

HomeworkTopicsDue dateFiles
Homework 1Policy gradients; variational diffusion modelsSep 17, 11:59 pmPDF · LaTeX source
Homework 2Language model alignment and compositionOct 10, 11:59 pmPDF · LaTeX source
Homework 3Diffusion model alignment and compositionOct 24, 11:59 pmPDF · LaTeX source
Homework 4Diffusion models for offline and online RLNov 21, 11:59 pmPDF · LaTeX source

Notes

  • Homework solutions must be written in LaTeX and submitted on Canvas. LaTeX source files are provided with each problem set.

  • Discussions with others are allowed, but all answers must be completed independently.

  • Any use of ChatGPT must be explicitly acknowledged, and students are responsible for the correctness of their answers.