Dream to Remember: Continual Steering of Generalist Robot Policies with World Models
cs.RO, cs.CV, cs.LG
Submitted: 2026-10-09
Updated: 2026-10-09
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control
- $\pi_\texttt{RL}$: Online RL Fine-tuning for Flow-based Vision-Language-Action Models
- DREAMSTEER: Latent World Models Can Steer VLA Policies During Deployment Without Any Finetuning
- DITTO: Offline Imitation Learning with World Models
- EXPO-FT: Sample-Efficient Reinforcement Learning Finetuning for Vision-Language-Action Models
- Continual Policy Consolidation for Lifelong Robot Learning
- VLA-RFT: Vision-Language-Action Reinforcement Fine-tuning with Verified Rewards in World Simulators
- CORAL: Scalable Multi-Task Robot Learning via LoRA Experts
- FlowDAgger: Human-in-the-Loop Adaptation of Generative Robot Policies in Latent Space
- GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
- $\pi^{*}_{0.6}$: a VLA That Learns From Experience
- Progressive Neural Networks
- Proximal Policy Optimization Algorithms
- World-Gymnast: Training Robots with Reinforcement Learning in a World Model
- Denoising Diffusion Implicit Models
- World-Env: Leveraging World Model as a Virtual Environment for VLA Post-Training
- Support-Constrained RL Enables Real-World Policy Improvement without Real-World Experience
- t-DGR: A Trajectory-Based Deep Generative Replay Method for Continual Learning in Decision Making
- Can Vision-Language-Action Models Learn from Real-World Data Continually without Forgetting?
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
- ARC: A Reasoning Recipe for Robot Foundation Models
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies