CFG-Policy: Unified Diffusion Policy Learning Framework via Classifier-free Guidance
cs.RO, cs.LG
Submitted: 2026-10-10
Updated: 2026-10-10
Project page: https://zhy-liu.github.io/cfg-policy-page
Sources
- Avoiding mode collapse in diffusion models fine-tuned with reinforcement learning
- $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control
- Exploration by Random Network Distillation
- DIME:Diffusion-Based Maximum Entropy Reinforcement Learning
- DiWA: Diffusion Policy Adaptation with World Models
- $\pi_\texttt{RL}$: Online RL Fine-tuning for Flow-based Vision-Language-Action Models
- RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation
- Sample-Efficient Diffusion-based Reinforcement Learning with Critic Guidance
- Maximum Entropy Reinforcement Learning with Diffusion Policy
- Diffusion Guidance Is a Controllable Policy Improvement Operator
- Behavior-Regularized Diffusion Policy Optimization for Offline Reinforcement Learning
- IDQL: Implicit Q-Learning as an Actor-Critic Method with Diffusion Policies
- Classifier-Free Diffusion Guidance
- GenPO++: Generative Policy Optimization with Jacobian-free Likelihood Ratios
- $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
- Flow Matching for Generative Modeling
- Sample from What You See: Visuomotor Policy Learning via Diffusion Bridge with Observation-Embedded Stochastic Differential Equation
- Reinforcement Fine-Tuning of Flow-Matching Policies for Vision-Language-Action Models
- Efficient Online Reinforcement Learning for Diffusion Policy
- GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models
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