ARS: Agentic Reward System for Robot Learning
cs.RO, cs.LG
Submitted: 2026-09-28
Updated: 2026-09-28
Code: https://github.com/midea-ai/ars
Terminology
Sources
- Qwen3-VL Technical Report
- Agentic Reward Modeling: Verifying GUI Agent via Progressive Trajectory-Grounded Interaction
- RoVer: Robot Reward Model as Test-Time Verifier for Vision-Language-Action Model
- DenseReward: Dense Reward Learning via Failure Synthesis for Robotic Manipulation
- ProcVLM: Learning Procedure-Grounded Progress Rewards for Robotic Manipulation
- PRM-as-a-Judge: A Dense Evaluation Paradigm for Fine-Grained Robotic Auditing
- RoboReward: General-Purpose Vision-Language Reward Models for Robotics
- Reward as An Agent for Embodied World Models
- $\pi^{*}_{0.6}$: a VLA That Learns From Experience
- SOLE-R1: Video-Language Reasoning as the Sole Reward for On-Robot Reinforcement Learning
- EV-WM: Event-Verified World Models for Long-Horizon Robotic Manipulation
- Large Reward Models: Generalizable Online Robot Reward Generation with Vision-Language Models
- Stage-Transition Dense Reward Modeling for Reinforcement Learning
- WARP-RM: A Warp-Augmented Relative Progress Reward Model for Data Curation
- UR-VC: Unsupervised Robotic Value Correction for Time-Derived Progress Proxies
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