VLA-MOPD: On-Policy Distillation for Multi-Task VLA Policy Consolidation
cs.RO, cs.CV
Submitted: 2026-11-09
Updated: 2026-11-09
Code: https://github.com/yxuan1234/VLA-MOPD
Terminology
Sources
- On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes
- $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control
- Fast-in-Slow: A Dual-System Foundation Model Unifying Fast Manipulation within Slow Reasoning
- $\pi_\texttt{RL}$: Online RL Fine-tuning for Flow-based Vision-Language-Action Models
- DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
- PRPO: Aligning Process Reward with Outcome Reward in Policy Optimization
- PaLM-E: An Embodied Multimodal Language Model
- DualVLA: Building a Generalizable Embodied Agent via Partial Decoupling of Reasoning and Action
- Flow-OPD: On-Policy Distillation for Flow Matching Models
- NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Tasks
- Residual Reinforcement Learning for Robot Control
- Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success
- Reward as An Agent for Embodied World Models
- DiffusionOPD: A Unified Perspective of On-Policy Distillation in Diffusion Models
- DSDyn-VLA: A Dual-Stream Dynamic Manipulation Framework with Motion Perception, Future Awareness, and Realtime Correction
- LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning
- HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model
- MOPD: Multi-Teacher On-Policy Distillation for Capability Integration in LLM Post-Training
- FAST: Efficient Action Tokenization for Vision-Language-Action Models
- $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation