Pith. sign in

REVIEW 13 cited by

ARPO:End-to-End Policy Optimization for GUI Agents with Experience Replay

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2505.16282 v1 pith:AA7GL5KX submitted 2025-05-22 cs.CV

ARPO:End-to-End Policy Optimization for GUI Agents with Experience Replay

classification cs.CV
keywords agentsarpooptimizationlearningpolicytrainingcomplexend-to-end
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Training large language models (LLMs) as interactive agents for controlling graphical user interfaces (GUIs) presents a unique challenge to optimize long-horizon action sequences with multimodal feedback from complex environments. While recent works have advanced multi-turn reinforcement learning (RL) for reasoning and tool-using capabilities in LLMs, their application to GUI-based agents remains relatively underexplored due to the difficulty of sparse rewards, delayed feedback, and high rollout costs. In this paper, we investigate end-to-end policy optimization for vision-language-based GUI agents with the aim of improving performance on complex, long-horizon computer tasks. We propose Agentic Replay Policy Optimization (ARPO), an end-to-end RL approach that augments Group Relative Policy Optimization (GRPO) with a replay buffer to reuse the successful experience across training iterations. To further stabilize the training process, we propose a task selection strategy that filters tasks based on baseline agent performance, allowing the agent to focus on learning from informative interactions. Additionally, we compare ARPO with offline preference optimization approaches, highlighting the advantages of policy-based methods in GUI environments. Experiments on the OSWorld benchmark demonstrate that ARPO achieves competitive results, establishing a new performance baseline for LLM-based GUI agents trained via reinforcement learning. Our findings underscore the effectiveness of reinforcement learning for training multi-turn, vision-language GUI agents capable of managing complex real-world UI interactions. Codes and models:https://github.com/dvlab-research/ARPO.git.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 13 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Learning to Explore: Scaling Agentic Reasoning via Exploration-Aware Policy Optimization

    cs.AI 2026-05 unverdicted novelty 7.0

    An exploration-aware RL framework lets LLM agents adaptively explore only under high uncertainty via variational rewards and action grouping, yielding consistent gains on text and GUI agent benchmarks.

  2. Faithful Mobile GUI Agents with Guided Advantage Estimator

    cs.AI 2026-05 unverdicted novelty 7.0

    Faithful-Agent raises Trap SR in GUI agents from 13.88% to 80.21% via faithfulness-oriented SFT and GuAE-enhanced RFT with consistency rewards while retaining general performance.

  3. Turing Test on Screen: A Benchmark for Mobile GUI Agent Humanization

    cs.AI 2026-02 unverdicted novelty 7.0

    The work creates a new benchmark for humanizing GUI agent touch dynamics via a MinMax detector-agent model, a mobile touch dataset, and methods showing agents can match human behavior without losing task performance.

  4. Learning with a Single Rollout via Monte Carlo Pass@k Critic

    cs.LG 2026-06 unverdicted novelty 6.0

    SR-PPO trains a Pass@k critic from single-rollout Monte Carlo outcomes to enable token-level advantage estimation in language model RL, yielding stable training and Pass@128 gains on math benchmarks.

  5. ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents

    cs.AI 2026-05 unverdicted novelty 6.0

    ToolCUA introduces a trajectory scaling pipeline and staged RL to optimize GUI-tool switching, reaching 46.85% accuracy on OSWorld-MCP for a 66% relative gain over baseline.

  6. Modularized Reinforcement Learning on LLMs: From MDP Creation to Exploration and Learning

    cs.LG 2026-06 unverdicted novelty 5.0

    Survey mapping RL techniques onto LLM training and highlighting gaps in value-based, off-policy, and bootstrapping methods.

  7. StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents

    cs.AI 2026-06 unverdicted novelty 5.0

    StainFlow proposes global entity stain tracking and local stain evidence linking modules to improve process rewards for GUI agents, reporting 3.2% relative gain in online RL success and 1.8% in judgment accuracy on An...

  8. RLVR without Ineffective Samples: Group Prioritized Off-Policy Optimization for LLM Reasoning

    cs.LG 2026-05 unverdicted novelty 5.0

    POPO uses recency-based prioritized group replay and decoupled off-policy optimization to avoid zero-variance ineffective samples in RLVR, accelerating LLM reasoning finetuning with fewer rollouts.

  9. Learning to Explore: Scaling Agentic Reasoning via Exploration-Aware Policy Optimization

    cs.AI 2026-05 unverdicted novelty 5.0

    An exploration-aware policy optimization method lets LLM agents explore selectively via a variational-inference reward and action grouping, yielding consistent gains on text and GUI agent benchmarks.

  10. On Training Large Language Models for Long-Horizon Tasks: An Empirical Study of Horizon Length

    cs.AI 2026-05 unverdicted novelty 5.0

    Longer action horizons bottleneck LLM agent training through instability, but training with reduced horizons stabilizes learning and enables better generalization to longer horizons.

  11. UI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement Learning

    cs.AI 2025-09 conditional novelty 5.0

    UI-TARS-2 reaches 88.2 on Online-Mind2Web, 47.5 on OSWorld, 50.6 on WindowsAgentArena, and 73.3 on AndroidWorld while attaining 59.8 mean normalized score on a 15-game suite through multi-turn RL and scalable data generation.

  12. DocArena: Turning Raw Documents into Controllable Training Environments for Document Search Agents

    cs.CV 2026-05 unverdicted novelty 4.0

    DocArena automates creation of multimodal document QA training data via MLLM-based structuring and cross-page reasoning pairs, yielding agents with top retrieval and QA performance in unified tests.

  13. Agentic Reasoning for Large Language Models

    cs.AI 2026-01 unverdicted novelty 4.0

    The survey structures agentic reasoning for LLMs into foundational, self-evolving, and collective multi-agent layers while distinguishing in-context orchestration from post-training optimization and reviewing applicat...