1 Part 1 · Imitation learning
Part 1 check
The quiz at the end of a lesson tests that lesson right after you read it. Here the tasks from all lessons of the part come shuffled: that shows what has stuck and whether the topics connect in your head. At the end you get a score per lesson and links to the sections worth revisiting.
1.1–1.2Behavior Cloning and DAgger: distribution shift and recovery data
1.3Multimodal actions: why regression drives into the kettle
1.4–1.5Diffusion models and Diffusion Policy
1.6–1.7The transformer and ACT: attention, tokens, action chunks
1.8Flow matching: straight paths from noise to data
✓ Result
Check result
What to revisit
➡️
Next: Part 2, spatial representations, language and planning. The policies of Part 1 know one task each and do not understand spoken commands. Part 2 is about tying actions to the geometry of the scene and to language: Transporter and CLIPort, PerAct and planners built on large language models.