SNS-Grasp: Semantic-guided Noise Scaling for Grasp Generation

SNS-Grasp is an intent-driven grasp generation framework that uses semantic information to guide robotic grasping. It explores how semantic signals can help robotic systems better understand the intended use of objects and produce more context-aware grasping behaviors.

This work was accepted to AAAI 2026.

Research Question

How can semantic information help robotic systems infer intent and generate more human-aligned grasps?

Current Status

TODO: Add paper link, project page, code, figures, full author list, and your specific contribution.