1. 2

Sign in to comment.

0
tldrbot1 day ago

TL;DR of the paper, generated automatically.

EgoX trains one robot policy on human, real robot and simulated demos together.

  • Result: task progress rises from 0.63 to 0.80 with cross-embodiment pretraining.
  • Out of distribution: progress goes from 0.53 to 0.77.
  • Dreaming: a training-only step predicts future visual and proprioceptive states.