Research map / Vision-language and multimodal / Robot action and navigation
Gripper-aware Vision Language Action Models
Hanyi Zhang, Zihong Luo, Tianyu Li, Khang Nguyen, Basu Hela, Shreyas Kumar, Ngoc Tran, Feng Dai, Charith Munasinghe, Jorge Queralta, Giovanni Toffetti, Khoa Vo and 7 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add gripper-aware tokenizer and dual MoA to pi0.5 | pi0.5 · MiGA sim average | success rate: 58.38 -> 66.00 | Improved |
| add gripper-aware tokenizer and MoA to pi0 | pi0 · MiGA sim average | success rate: 38.13 -> 48.13 | Improved |
| add GVLA to pi0 on flat tasks | pi0 · MiGA sim flat | success rate: 30.00 -> 27.50 | No clear effect |
| add GVLA to pi0.5 on flat tasks | pi0.5 · MiGA sim flat | success rate: 52.50 -> 53.00 | No clear effect |
| add GVLA to pi0.5 on stacked tasks | pi0.5 · MiGA sim stacked | success rate: 71.00 -> 76.00 | Improved |
| use multi-level tokenizer vs MLP embeddings | GVLA · tokenizer comparison | prediction error: 0.053 -> 0.032 | Improved |
| use multi-level tokenizer vs VQ-VAE | GVLA · tokenizer comparison | CAPD: 0.70 -> 1.34 | Improved |
| add platform token (ablation) | GVLA · unseen-gripper adaptation | adaptation success: 0.54 -> 0.92 | Improved |
| add type token (ablation) | GVLA · unseen-gripper adaptation | adaptation success: 0.86 -> 0.92 | Improved |
| add instance token (ablation) | GVLA · unseen-gripper adaptation | adaptation success: 0.90 -> 0.92 | Improved |
| add dual MoA (ablation) | GVLA · unseen-gripper adaptation | adaptation success: 0.52 -> 0.92 | Improved |
| add dual MoA (ablation) | GVLA · MiGA | prediction error: 0.037 -> 0.032 | Improved |
Closest papers by meaning
- ForceVLA: Enhancing VLA Models with a Force-aware MoE for Contact-rich Manipulation (NeurIPS 2025 · similarity 0.90)
- Exploring the Limits of Vision-Language-Action Manipulation in Cross-task Generalization (NeurIPS 2025 · similarity 0.89)
- Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.89)
- VLA-OS: Structuring and Dissecting Planning Representations and Paradigms in Vision-Language-Action Models (NeurIPS 2025 · similarity 0.89)
- CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models (CVPR 2025 · similarity 0.89)
- RealVLG-R1: A Large-Scale Real-World Visual-Language Grounding Benchmark for Robotic Perception and Manipulation (CVPR 2026 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
