Metraipolis › archive › dc_14c13a772ff8
PastForward: Faster On-Device GUI Agents via Computational Experience Reuse — arXiv:2609.32166
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.32166
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T20:55:08.714Z
- SHA-256 of the archived text
33921a3c440ad63675c58450b1808b90d3d28a298b0d5f2e3cb7773ad3583878- Stable URL of this page
https://metraipolis.com/documento/dc_14c13a772ff8
How to cite
PastForward: Faster On-Device GUI Agents via Computational Experience Reuse — arXiv:2609.32166 — arXiv (en), https://arxiv.org/abs/2609.32166 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_14c13a772ff8 (Metraipolis archive dc_14c13a772ff8, sha256 33921a3c440ad63675c58450b1808b90d3d28a298b0d5f2e3cb7773ad3583878)
BibTeX
@misc{dc14c13a772ff8,
title = {PastForward: Faster On-Device GUI Agents via Computational Experience Reuse — arXiv:2609.32166},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.32166},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_14c13a772ff8; sha256 33921a3c440ad63675c58450b1808b90d3d28a298b0d5f2e3cb7773ad3583878},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32166. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Running GUI agents on edge devices can keep sensitive screens and interaction histories local, but the computational cost of inference at every action step makes deployment challenging. Existing GUI agent systems either perform full vision-language model (VLM) inference at each action step or reuse coarse-grained knowledge matched to prior tasks. However, dynamic mobile environments and user tasks make it difficult to fully utilize prior task executions without additional fine-tuning or task-specific offline exploration. To address this challenge, we present PastForward, a system that accelerates GUI agents through validated, fine-grained reuse of computational experience accumulated during ordinary task execution. During decoding, PastForward retrieves prior output sequences as device-adaptive multi-token proposals and verifies them in a single VLM forward pass. Across action steps, it uses prior GUI transitions to begin next-step inference while the device executes the current action, retains the early computation only when the predicted screen matches the observed screen, and carries reusable KV states forward. We evaluate PastForward on AndroidWorld workloads derived from real mobile usage patterns using multiple VLM backbones across server and edge platforms. On device, PastForward achieves action-step latency speedups of 1.63-2.36$\times$ while maintaining task success rates. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.32166 — 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 2026-10-07T20:29:00.718Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).