Stanford's Jeannette Bohg to present research on fixing VLA and diffusion policy failures in contact-rich robotic tasks · Digg