# Critic in the Loop: A Tri-System VLA Framework for Robust Long-Horizon Manipulation

arXiv: https://arxiv.org/abs/2603.05185 · HTML: https://mlsanity.com/paper/2603.05185/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Efficient VLA inference and adaptation, one of 13 directions in Vision-language-action models.
- Builds on 18 of its 23 references that are in the map.
- First posted on arXiv 5 Mar 2026.

## Placed in

- Vision-language-action models / [Efficient VLA inference and adaptation](https://mlsanity.com/vla/efficient-vla-inference-and-adaptation/)

## Builds on (most cited first, 18 of 18)

- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [π0.5: a Vision-Language-Action Model with Open-World Generalization](https://mlsanity.com/paper/2504.16054/) (2025)
- [GR00T N1: An Open Foundation Model for Generalist Humanoid Robots](https://mlsanity.com/paper/2503.14734/) (2025)
- [PaliGemma: A versatile 3B VLM for transfer](https://mlsanity.com/paper/2407.07726/) (2024)
- [π*0.6: a VLA That Learns From Experience](https://mlsanity.com/paper/2511.14759/) (2025)
- [Hi Robot: Open-Ended Instruction Following with Hierarchical Vision-Language-Action Models](https://mlsanity.com/paper/2502.19417/) (2025)
- [HAMSTER: Hierarchical Action Models For Open-World Robot Manipulation](https://mlsanity.com/paper/2502.05485/) (2025)
- [OneTwoVLA: A Unified Vision-Language-Action Model with Adaptive Reasoning](https://mlsanity.com/paper/2505.11917/) (2025)
- [OpenHelix: A Short Survey, Empirical Analysis, and Open-Source Dual-System VLA Model for Robotic Manipulation](https://mlsanity.com/paper/2505.03912/) (2025)
- [RoboReward: General-Purpose Vision-Language Reward Models for Robotics](https://mlsanity.com/paper/2601.00675/) (2026)
- [GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulation](https://mlsanity.com/paper/2512.01801/) (2025)
- [A Dual Process VLA: Efficient Robotic Manipulation Leveraging VLM](https://mlsanity.com/paper/2410.15549/) (2024)
- [Robo-Dopamine: General Process Reward Modeling for High-Precision Robotic Manipulation](https://mlsanity.com/paper/2512.23703/) (2025)
- [RISE: Self-Improving Robot Policy with Compositional World Model](https://mlsanity.com/paper/2602.11075/) (2026)
- [Logic-in-Frames: Dynamic Keyframe Search via Visual Semantic-Logical Verification for Long Video Understanding](https://mlsanity.com/paper/2503.13139/) (2025)
- [K-frames: Scene-Driven Any-k Keyframe Selection for long video understanding](https://mlsanity.com/paper/2510.13891/) (2025)
- [VAMOS: A Hierarchical Vision-Language-Action Model for Capability-Modulated and Steerable Navigation](https://mlsanity.com/paper/2510.20818/) (2025)
- [FoundationMotion: Auto-Labeling and Reasoning about Spatial Movement in Videos](https://mlsanity.com/paper/2512.10927/) (2025)

## Built on by (most cited first, 4 of 4)

- [Goal2Skill: Long-Horizon Manipulation with Adaptive Planning and Reflection](https://mlsanity.com/paper/2604.13942/) (2026)
- [Foresight: Failure Detection for Long-Horizon Robotic Manipulation with Action-Conditioned World Model Latents](https://mlsanity.com/paper/2606.23085/) (2026)
- [DriveJudge: Rethinking Autonomous Driving Evaluation with Vision-Language Models](https://mlsanity.com/paper/2606.17362/) (2026)
- [Skill-Space Shooting for Autonomous Robot Policy Improvement](https://mlsanity.com/paper/2609.38178/) (2026)

## Abstract (arXiv)

Balancing high-level semantic reasoning with low-level reactive control remains a core challenge in visual robotic manipulation. While Vision-Language Models (VLMs) excel at cognitive planning, their inference latency precludes real-time execution. Conversely, fast Vision-Language-Action (VLA) models often lack the semantic depth required for complex, long-horizon tasks. To bridge this gap, we introduce Critic in the Loop, an adaptive hierarchical framework driven by dynamic VLM-Expert scheduling. At its core is a bionic Tri-System architecture comprising a VLM brain for global reasoning, a VLA cerebellum for reactive execution, and a lightweight visual Critic. By continuously monitoring the workspace, the Critic dynamically routes control authority. It sustains rapid closed-loop execution via the VLA for routine subtasks, and adaptively triggers the VLM for replanning upon detecting execution anomalies such as task stagnation or failures. Furthermore, our architecture seamlessly integrates human-inspired rules to intuitively break infinite retry loops. This visually-grounded scheduling minimizes expensive VLM queries, while substantially enhancing system robustness and autonomy in out-of-distribution (OOD) scenarios. Comprehensive experiments on challenging, long-horizon manipulation benchmarks reveal that our approach achieves state-of-the-art performance.