This schedule is tentative and subject to change.
| WEEK | TOPIC | DATE | LECTURE/DISCUSSION |
|---|---|---|---|
| 1 | Intro and RL-post-training of LLMs | 09/24 | Introduction to course, syllabus. RL basics - Lecture (slides) (recording) |
| 09/26 | RL post-training history and RLHF intro - Lecture (slides) (recording) | ||
| 2 | RL from Human Feedback (RLHF) (for LLMs) | 10/1 | RLHF - Discussion |
| 10/3 | Personalized RLHF - Lecture (slides) (recording) | ||
| 3 | RLVR and reasoning, embodied instruction following | 10/8 | RLVR, reasoning, and technical aspects of RL post-training - Discussion |
| 10/10 | Embodied instruction following - Discussion | ||
| 4 | Learning from humans beyond LLMs | 10/15 | Inverse RL and other ways to learn from humans - Lecture |
| 10/17 | PbRL / robot learning from human feedback - Discussion | ||
| 5 | Multi-agent RL | 10/22 | Deep multi-agent RL intro - Lecture |
| 10/24 | State-of-the-art MARL, theory, fun MARL papers, and social skills like theory of Mind - Discussion | ||
| 6 | Human-AI coordination | 10/29 | Zero-shot coordination with humans - Lecture |
| 10/31 | Human-agent coordination and population-based training - Discussion | ||
| 7 | Emergent complexity | 11/5 | Emergent Complexity - Lecture |
| 11/7 | Emergent Complexity and Open-Endedness - Discussion | ||
| 8 | Social learning | 11/12 | Lecture |
| 11/14 | Social Learning - Discussion | ||
| 9 | Multi-agent RL for LLMs | 11/19 | Multi-agent RL for LLM - Lecture |
| 11/21 | MARL for LLMs, Multi-agent LLM systems, Red-teaming - Lecture | ||
| 10 | Extra topics | 11/26 | Bonus Lecture |
| 11/28 | No class | ||
| 11 | Final project presentations | 12/3 | Final project presentations |
| 12/5 | Final project presentations |