Research map / Vision-language and multimodal / Robot action and navigation
LIBERO-Plus: A Progressive Robustness Benchmark for Visual-Language-Action Models
Senyu Fei, Siyin Wang, Junhao Shi, Zihao Dai, Jikun Cai, Pengfang Qian, Li Ji, Xinzhe He, Shiduo Zhang, Zhaoye Fei, Jinlan Fu, Jingjing Gong and 1 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 |
|---|---|---|---|
| use wrist camera in addition to third-person view | OpenVLA-OFT vs OpenVLA-OFT w · LIBERO-Plus, Camera perturbation | success rate (%): 10.4 -> 56.4 | Improved |
| add confounding objects to layout | pi0, pi0-fast, RIPT-VLA, UniVLA, WorldVLA · LIBERO-Plus, Layout | success rate (%) | No clear effect |
| displace the target object placement and pose | VLA models · LIBERO-Plus, Layout | success rate (%) | Hurt |
| mask third-person view only (keep wrist camera) (ablation) | OpenVLA-OFT · LIBERO, light ablation | success rate (%): 97.1 -> 43.6 | Hurt |
| mask both camera views (ablation) | OpenVLA-OFT · LIBERO, light ablation | success rate (%): 97.1 -> 0.0 | Hurt |
| add light perturbation to models with wrist camera | OpenVLA-OFT · LIBERO-Plus, Light | success rate (%): 97.1 -> 88.7 | No clear effect |
| perturb language instructions | OpenVLA-OFT · LIBERO-Plus, Language | success rate (%): 97.1 -> 79.5 | No clear effect |
| remove the language instruction (blank) (ablation) | OpenVLA-OFT · LIBERO object suite | success rate (%) | No clear effect |
| remove the language instruction (blank) (ablation) | OpenVLA-OFT · LIBERO long suite | success rate (%) | Hurt |
| replace the target object in instruction and goal (ablation) | OpenVLA-OFT · LIBERO object suite (modified target) | success rate (%) | Hurt |
| apply pairwise combined perturbations | OpenVLA-OFT · LIBERO-Plus pairwise | joint success probability: 18.93 -> 18.29 (-0.64) | Hurt |
| fine-tune on generated generalization dataset (20,000+ trajectories) | OpenVLA-OFT · LIBERO-Plus, Camera | success rate (%): 55.6 -> 92.8 | Improved |
Closest papers by meaning
- VLA-REPLICA: A Low-Cost, Reproducible Benchmark for Real-World Evaluation of Vision-Language-Action Models (NeurIPS 2026 · similarity 0.91)
- ActionUNet: Improving Robustness of VLA Models with Efficient Multi-scale Fine-tuning (NeurIPS 2026 · similarity 0.90)
- PriorVLA: Prior-Preserving Adaptation for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
- Exploring the Limits of Vision-Language-Action Manipulation in Cross-task Generalization (NeurIPS 2025 · similarity 0.89)
- CofactVLA: Deconfounding Vision-Language-Action Models via Counterfactual Intervention (NeurIPS 2026 · similarity 0.89)
- ATI-VLA: Action-Centric Predictive Vision–Language–Action Models via Actionable Alignment Then Adaptive Injection (NeurIPS 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.
