LIT’s approach to teaching goal-directed motion before adding vision
The post argues that learning movement without images can avoid visual shortcuts. Later vision training is meant to support the action expert’s existing spatial understanding.
TLDR
A post explaining LIT describes two training stages. First, an “action expert” learns to move toward a spatial goal without images. Then, vision is introduced through a latent interface trained to recover that same goal. The author’s rationale is to establish goal-directed movement first, then encourage visual information to support what matters for the action.
Combined views
15.4K
3 Sources, first seen ago
