Michael Stark

RT-Sketch: Goal-Conditioned Imitation Learning from Hand-Drawn Sketches

Generating Robot Policy Code for High-Precision and Contact-Rich Manipulation Tasks