Robotics & Embodied AI

Vision-language-action models, action tokenisation, diffusion policies and sim-to-real.

10concepts
122flashcards
80minutes of reading
  1. 01 Action Tokenisation and Representation How continuous robot actions become discrete tokens an autoregressive transformer can emit, from per-dimension binning to frequency-space compression. intermediate 7m 6 cards
  2. 02 Evaluating Robot Policies Why a reported 70 percent success rate on a robot task usually carries an error bar of plus or minus 15 points, and what protocol changes make two policies actually comparable. intermediate 8m 18 cards
  3. 03 Imitation Learning and Diffusion Policies Why cloning a demonstrator's actions drifts into unseen states, and how generative action models such as diffusion policies and action chunking control the drift. intermediate 8m 7 cards
  4. 04 Massively Parallel Simulation and Sim-to-Real Locomotion How putting the physics engine on the GPU turned weeks of legged-robot RL into minutes, and why domain randomisation is the only reason those policies survive contact with a real floor. intermediate 8m 18 cards
  5. 05 Sim-to-Real Transfer Why robot policies are trained in simulation, why they break on real hardware, and how domain randomisation closes the reality gap by making the real world look like one more random draw. intermediate 7m 7 cards
  6. 06 Cross-Embodiment Transfer Why robot data has never had an ImageNet, how Open X-Embodiment pooled 22 robot types into one training set, and what has to be true for data from one arm to help a different arm. advanced 9m 18 cards
  7. 07 Flow-Matching Action Experts Why binning continuous robot actions into text tokens caps control frequency, and how attaching a flow-matching action head to a vision-language model reaches 50 Hz dexterous control. advanced 9m 18 cards
  8. 08 RT-2 and Web-Scale Robot Learning RT-2 co-trains one transformer on internet vision-language data and robot trajectories by encoding actions as text tokens, transferring semantic web knowledge into robotic control. advanced 7m 6 cards
  9. 09 Vision-Language-Action Models Turning a pretrained vision-language model into a robot policy that maps camera images plus a language instruction to motor actions, so the robot inherits web-scale semantic knowledge it could never learn from robot data alone. advanced 8m 6 cards
  10. 10 World Models for Control How learning a latent dynamics model lets an agent train inside its own imagination, why that is the answer to robotics' sample-efficiency problem, and where the learned simulator's errors get exploited. advanced 9m 18 cards