RealGUINoise: An Interactive Cross-Platform Benchmark for GUI Agent Robustness under Real-World Interface Noise
4.00T1 sourcearXiv cs.HC
arXiv cs.HCoriginal source ↗
Source record
Published by arXiv cs.HC (T1 source). The original is at https://arxiv.org/abs/2609.38184.
Pipeline notes
The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.
SummaryRealGUINoise is a cross-platform, interactive benchmark covering 42 noise types across web, desktop, and mobile, integrating 7 GUI agent frameworks. Experiments show real-world interface noise degrades task performance, redirects action trajectories, and increases unsafe behavior.
Why it mattersConcrete evaluation framework and reproducible findings on how everyday interface noise breaks GUI agents, useful for anyone building or selecting computer-using agents.
Cited by
No citations on record.