Flow control steers VLA models with real-time user inputs to achieve higher success rates and faster task completion while maintaining action quality.
Inference-time policy steering through human interactions
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.RO 3roles
background 1polarities
background 1representative citing papers
DexVLA combines a scaled diffusion action expert with embodiment curriculum learning to achieve better generalization and performance than prior VLA models on diverse robot hardware and long-horizon tasks.
citing papers explorer
-
Flow Control: Steering Vision-Language-Action Models with Simple Real-Time Inputs
Flow control steers VLA models with real-time user inputs to achieve higher success rates and faster task completion while maintaining action quality.
- You've Got a Golden Ticket: Improving Generative Robot Policies With A Single Noise Vector