Metraipolis › archive › dc_bb206835d379

STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI Agents — arXiv:2610.20390

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.20390
License (as archived)
CC0-1.0
Archived
2026-10-15T21:57:15.093Z
SHA-256 of the archived text
ad8e5c27fabad04c9f58efe696fcab1a52f862862ca43664c30b6e3ede7ffa74
Stable URL of this page
https://metraipolis.com/documento/dc_bb206835d379

How to cite

STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI Agents — arXiv:2610.20390 — arXiv (en), https://arxiv.org/abs/2610.20390 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bb206835d379 (Metraipolis archive dc_bb206835d379, sha256 ad8e5c27fabad04c9f58efe696fcab1a52f862862ca43664c30b6e3ede7ffa74)
BibTeX
@misc{dcbb206835d379,
  title = {STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI Agents — arXiv:2610.20390},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.20390},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bb206835d379; sha256 ad8e5c27fabad04c9f58efe696fcab1a52f862862ca43664c30b6e3ede7ffa74},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20390. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

With the rapid development of vision-language models (VLMs), Web-GUI agents have achieved promising performance in interacting with real-world interfaces through screenshot-based grounding. However, high-resolution screenshots and multi-turn histories introduce excessive visual tokens, causing substantial inference overhead. Existing token pruning methods, mainly designed for natural images, often select tokens independently based on importance scores, which can fragment the local visual structures required for recognizing small GUI elements such as buttons, icons, and hyperlinks. In this work, we propose STP-GUI (Spatially-aware Token Pruning for GUI agents), a query-guided token pruning framework that preserves fine-grained instruction-conditioned token scoring while performing spatially coherent window-level selection under aggressive token reduction. A lightweight selector is optimized through policy learning while the VLM backbone remains frozen. Experiments on Online-Mind2Web show that \papername achieves a favorable efficiency-performance tradeoff under low token budgets. At 21% token retention, it preserves 77.7% of full-token performance while reducing inference latency by 31% and FLOPs by 57%. Results on Android in the Wild further suggest that the method transfers to mobile GUI environments.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.20390 — 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 2026-10-15T21:38:29.537Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).