Research map / Vision-language and multimodal / Robot action and navigation
villa-X: Enhancing Latent Action Modeling in Vision-Language-Action Models
Xiaoyu Chen, Hangxing Wei, Pushi Zhang, Chuheng Zhang, Kaixin Wang, Yanjiang Guo, Rushuai Yang, Yucen Wang, Xinquan Xiao, Li Zhao, Jianyu Chen, Jiang Bian
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 proprioceptive FDM to the latent action model (ablation) | villa-X LAM · SIMPLER (small pretrain mix) Google robot | avg success rate (%): 57.4 -> 58.5 | Improved |
| add proprioceptive FDM to the latent action model (ablation) | villa-X LAM · SIMPLER (small pretrain mix) WidowX | avg success rate (%): 32.3 -> 40.8 | Improved |
| pretrain with latent actions (ablation) | villa-X · SIMPLER (small pretrain mix) Google robot | avg success rate (%): 35.0 -> 58.5 | Improved |
| pretrain with latent actions (ablation) | villa-X · SIMPLER (small pretrain mix) WidowX | avg success rate (%): 33.1 -> 40.8 | Improved |
| joint diffusion latent expert instead of LAPA-style two-stage (ablation) | villa-X · SIMPLER (small pretrain mix) Google robot | avg success rate (%): 43.8 -> 58.5 | Improved |
| joint diffusion latent expert instead of Go-1-style planner (ablation) | villa-X · SIMPLER (small pretrain mix) WidowX | avg success rate (%): 14.8 -> 40.8 | Improved |
| add latent action expert (ablation) | villa-X · SIMPLER Google robot (full pretrain) | avg success rate (%): 36.5 -> 77.7 | Improved |
| add latent action expert (ablation) | villa-X · LIBERO | avg success rate (%): 81.9 -> 90.1 | Improved |
| add stochastic attention masking from robot to latent actions (ablation) | villa-X policy · SIMPLER (small pretrain mix) Google robot | avg success rate (%): 53.2 -> 58.5 | Improved |
| add stochastic attention masking from robot to latent actions (ablation) | villa-X policy · SIMPLER (small pretrain mix) WidowX | avg success rate (%): 34.0 -> 40.8 | Improved |
| add embodiment context to policy (ablation) | villa-X policy · SIMPLER (small pretrain mix) Google robot | avg success rate (%): 49.1 -> 58.5 | Improved |
| add embodiment context to policy (ablation) | villa-X policy · SIMPLER (small pretrain mix) WidowX | avg success rate (%): 38.5 -> 40.8 | Mixed |
Later papers that compared with this method
5 later accepted papers report Villa-x as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- UniJEPA: Enhancing Robot Policy via Unified Continuous and Discrete Representation Learning (ICML 2026 · better on 2 of 2 shared results)
- LARA: Latent Action Representation Alignment for Vision-Language-Action Models (ICML 2026 · better on 2 of 2 shared results)
- RotVLA: Rotational Latent Action for Vision-Language-Action Model (NeurIPS 2026 · better on 2 of 2 shared results)
- Chain of World: World Model Thinking in Latent Motion (CVPR 2026 · better on 2 of 2 shared results)
- VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Model (ECCV 2026 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- NS-VLA: Towards Neuro-Symbolic Vision-Language-Action Models (NeurIPS 2026 · similarity 0.92)
- Latent Action Pretraining from Videos (ICLR 2025 · similarity 0.92)
- RotVLA: Rotational Latent Action for Vision-Language-Action Model (NeurIPS 2026 · similarity 0.91)
- Exploring the Limits of Vision-Language-Action Manipulation in Cross-task Generalization (NeurIPS 2025 · similarity 0.91)
- Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.91)
- PriorVLA: Prior-Preserving Adaptation for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
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.
