Dense to MoE Adaptation for Compact Vision Language Action Policies
cs.RO
Submitted: 2026-09-15
Updated: 2026-09-18
Terminology
Sources
- RT-1: Robotics Transformer for Real-World Control at Scale
- RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation
- HiMoE-VLA: Hierarchical Mixture-of-Experts for Generalist Vision-Language-Action Policies
- Mixtral of Experts
- OpenVLA: An Open-Source Vision-Language-Action Model
- Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints
- GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding
- Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Models
- Flow Matching for Generative Modeling
- MLPMoE: Zero-Shot Architectural Metamorphosis of Dense LLM MLPs into Static Mixture-of-Experts
- Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
- Expertise need not monopolize: Action-Specialized Mixture of Experts for Vision-Language-Action Learning
- Auxiliary-Loss-Free Load Balancing Strategy for Mixture-of-Experts
Related papers
- FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
- MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
- RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies
- Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving