Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better

NeurIPSSpotlight2025

Authors: Danny Driess, Jost Tobias Springenberg, brian ichter, LILI YU, Adrian Li-Bell, Karl Pertsch, Allen Z. Ren, Homer Walke, Quan Vuong, Lucy Xiaoyang Shi, Sergey Levine
Venue: NeurIPS 2025
Track: Spotlight

TL;DR

Vision-language-action (VLA) models provide a powerful approach to training control policies for physical systems, such as robots, by combining end-to-end learning with transfer of semantic knowledge…

Opening excerpt from the authors’ abstract. source

Read the paper

OpenReview

Topics

vision-language control

← All NeurIPS 2025 Spotlight papers · Browse the whole archive