Practical Vision-Language-Action Models

This line of work studies the gap between benchmark performance and practical robot deployment, with an emphasis on robust evaluation and improved VLA baselines.