# Reinforcement Learning for Computer-Use Agents with Autonomous Evaluation

arXiv: https://arxiv.org/abs/2606.24515 · HTML: https://mlsanity.com/paper/2606.24515/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in GUI and mobile agents, one of 16 directions in LLM agents and tool use.
- Builds on 22 of its 29 references that are in the map.
- First posted on arXiv 23 Jun 2026.

## Placed in

- LLM agents and tool use / [GUI and mobile agents](https://mlsanity.com/agents/gui-and-mobile-agents/)
- LLM alignment / RLHF / preference optimisation / [Process reward models for reasoning](https://mlsanity.com/alignment/process-reward-models-for-reasoning/)

## Builds on (most cited first, 22 of 22)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [WebArena: A Realistic Web Environment for Building Autonomous Agents](https://mlsanity.com/paper/2307.13854/) (2023)
- [UI-TARS: Pioneering Automated GUI Interaction with Native Agents](https://mlsanity.com/paper/2501.12326/) (2025)
- [WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learning](https://mlsanity.com/paper/2411.02337/) (2024)
- [Windows Agent Arena: Evaluating Multi-Modal OS Agents at Scale](https://mlsanity.com/paper/2409.08264/) (2024)
- [Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes](https://mlsanity.com/paper/2408.03539/) (2024)
- [OmniACT: A Dataset and Benchmark for Enabling Multimodal Generalist Autonomous Agents for Desktop and Web](https://mlsanity.com/paper/2402.17553/) (2024)
- [Reinforcement Learning with Perturbed Rewards](https://mlsanity.com/paper/1810.01032/) (2018)
- [SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience](https://mlsanity.com/paper/2508.04700/) (2025)
- [InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection](https://mlsanity.com/paper/2501.04575/) (2025)
- [Pgx: Hardware-Accelerated Parallel Game Simulators for Reinforcement Learning](https://mlsanity.com/paper/2303.17503/) (2023)
- [ComputerRL: Scaling End-to-End Online Reinforcement Learning for Computer Use Agents](https://mlsanity.com/paper/2508.14040/) (2025)
- [Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers](https://mlsanity.com/paper/2510.00915/) (2025)
- [GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding](https://mlsanity.com/paper/2406.10819/) (2024)
- [A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions](https://mlsanity.com/paper/2501.16150/) (2025)
- [macOSWorld: A Multilingual Interactive Benchmark for GUI Agents](https://mlsanity.com/paper/2506.04135/) (2025)
- [CUARewardBench: A Benchmark for Evaluating Reward Models on Computer-using Agent](https://mlsanity.com/paper/2510.18596/) (2025)
- [The Art of Building Verifiers for Computer Use Agents](https://mlsanity.com/paper/2604.06240/) (2026)
- [CANDERE-COACH: Reinforcement Learning from Noisy Feedback](https://mlsanity.com/paper/2409.15521/) (2024)
- ["Are We Done Yet?": A Vision-Based Judge for Autonomous Task Completion of Computer Use Agents](https://mlsanity.com/paper/2511.20067/) (2025)
- [CUAAudit: Meta-Evaluation of Vision-Language Models as Auditors of Autonomous Computer-Use Agents](https://mlsanity.com/paper/2603.10577/) (2026)

## Built on by (most cited first, 1 of 1)

- [SeekJudge: A Practical Reward Framework for Reinforcement Learning in Computer-Use Agents](https://mlsanity.com/paper/2607.23263/) (2026)

## Abstract (arXiv)

Computer-Use Agents (CUAs) execute high-level user goals by perceiving and acting directly within graphical user interfaces. However, reinforcement learning for CUAs remains difficult because open-ended desktop environments rarely provide scalable, machine-readable reward signals: task success is often visually grounded and hard to specify with handcrafted reward functions or dense manual labels. We propose an RL fine-tuning framework that uses autonomous vision-language evaluation as a scalable supervision signal for GUI agents. Given a final screenshot and the original instruction, a Vision-Language Model judges task completion and provides terminal feedback without task-specific heuristics or manual labels during policy optimization. Because autonomous evaluators are imperfect, we model their feedback as a noisy binary reward channel and derive a noise-corrected reward estimator for Proximal Policy Optimization. Experiments across macOSWorld, Windows Agent Arena, and OSWorld show that corrected evaluator rewards outperform both zero-shot baselines and raw evaluator rewards, improving success rates by an average of 12.6 percentage points over zero-shot performance and 5.1 points over raw evaluator fine-tuning. These results suggest that autonomous evaluation can serve as a practical reward signal for RL in GUI environments when evaluator noise is explicitly modeled and corrected.