Metraipolis › archive › dc_9ca507acca92

GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis — arXiv:2608.22847

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.22847
License (as archived)
CC0-1.0
Archived
2026-10-18T07:41:40.020Z
SHA-256 of the archived text
5372f5759970acbc787601a85782dd16991edb93f57509238d7ae5e2b0f14e4a
Stable URL of this page
https://metraipolis.com/documento/dc_9ca507acca92

How to cite

GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis — arXiv:2608.22847 — arXiv (en), https://arxiv.org/abs/2608.22847 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9ca507acca92 (Metraipolis archive dc_9ca507acca92, sha256 5372f5759970acbc787601a85782dd16991edb93f57509238d7ae5e2b0f14e4a)
BibTeX
@misc{dc9ca507acca92,
  title = {GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis — arXiv:2608.22847},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.22847},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9ca507acca92; sha256 5372f5759970acbc787601a85782dd16991edb93f57509238d7ae5e2b0f14e4a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.22847. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-Language Models (VLMs) based GUI agents stand to benefit significantly from online reinforcement learning (RL). However, their training is bottlenecked by two fundamental issues: current data synthesis methods for GUI Agents rely on specific environments and struggle to generate diverse data, while existing evaluators either suffer from limited scalability or provide inaccurate and unreliable reward signals. To overcome these challenges, we introduce GSAR (Goal-State-Anchor Reward), a RL reward framework that supports scalable task generation and delivers reliable reward signals for stable and efficient policy optimization. Our approach features self-evolving data synthesis, which produces multiple environments through task execution and generates diverse tasks and goal states. Complementing this, a state-anchor mechanism automatically annotates task-relevant UI elements in successful goal states as reference anchors. During RL training, these reference anchors provide accurate, scalable reward signals that substantially enhance efficiency. Extensive evaluations demonstrate that our framework achieves over 90% accuracy on offline trajectory verification and performs closest to rule-based methods. Furthermore, agents trained using our reward framework exhibit strong performance on both AndroidWorld and our constructed benchmark, establishing a scalable approach for GUI agent training.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.22847 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-18T07:37:52.747Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).