Research map / Vision-language and multimodal / Robot action and navigation
Learning to See and Act: Task-Aware Virtual View Exploration for Robotic Manipulation
Yongjie Bai, Zhouxia Wang, Yang Liu, Kaijun Luo, Yifan Wen, Mingtong Dai, Weixing Chen, Ziliang Chen, Lingbo Liu, Guanbin Li, Liang Lin
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 |
|---|---|---|---|
| replace fixed viewpoints with MVEP-explored virtual views (ablation) | TVVE · RLBench multi-view (18 tasks) | avg success rate (%): 83.3 -> 86.6 | Improved |
| use learned viewpoints instead of random viewpoints (ablation) | TVVE · RLBench multi-view (18 tasks) | avg success rate (%): 8.9 -> 86.6 | Improved |
| add TaskMoE (ablation) | TVVE · RLBench multi-view (18 tasks) | avg success rate (%): 85.6 -> 86.6 | Improved |
| add TaskMoE (seen tasks) (ablation) | TVVE · RLBench 5-task subset | avg success rate (%): 72.0 -> 80.8 | Improved |
| add TaskMoE (unseen task) (ablation) | TVVE · RLBench Open Drawer | success rate (%): 60.0 -> 72.0 | Improved |
| increase number of views K from 2 to 4 (ablation) | TVVE · RLBench 5-task subset | avg success rate (%): 27.2 -> 55.2 | Improved |
| tighten camera radial range to 0.90-1.04 m (ablation) | TVVE · RLBench 5-task subset | avg success rate (%): 49.6 -> 56.0 | Improved |
| loosen camera radial range to 0.60-1.56 m (ablation) | TVVE · RLBench 5-task subset | avg success rate (%): 49.6 -> 48.8 | Hurt |
| use TVVE over ARP on RLBench multi-view | TVVE · RLBench multi-view (18 tasks) | avg success rate (%): 81.6 -> 86.6 | Improved |
| use TVVE over ARP under occlusion/OOD | TVVE · RLBench-OG | avg success rate (%): 63.7 -> 67.0 | Improved |
| use TVVE over ARP under camera pose change | TVVE · RLBench-OG Camera Pose | success rate (%): 73.2 vs ARP 69.7 vs RVT2 74.0 | Mixed |
| use TVVE over Diffusion Policy on real robot | TVVE · Real-world Dobot | avg success rate (%): 68.0 -> 88.0 | Improved |
Closest papers by meaning
- Seeing Across Views: Benchmarking Spatial Reasoning of Vision-Language Models in Robotic Scenes (ICLR 2026 · similarity 0.89)
- Localizing, Structuring, and Rendering: Bridging 3D and 2D Vision-Language-Action Models for Robotic Manipulation (CVPR 2026 · similarity 0.89)
- Exploring the Limits of Vision-Language-Action Manipulation in Cross-task Generalization (NeurIPS 2025 · similarity 0.89)
- Object-Centric Prompt-Driven Vision-Language-Action Model for Robotic Manipulation (CVPR 2025 · similarity 0.89)
- VersaCamVLA: Camera-Configurable VLA Policies for Robotic Manipulation (NeurIPS 2026 · similarity 0.89)
- Robotic Visual Instruction (CVPR 2025 · similarity 0.88)
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.
