Papers
arxiv:2611.22722

Practice in Text, Act in Pixels: Correcting Where Perception Fails

Published on Nov 17
Authors:
,

Abstract

Aligning a vision-language screen reader on a small set of labeled screenshots yields higher GUI task success than training the policy on those screenshots, and one aligned reader generalizes across text-trained policies.

GUI agents can learn cheaply from a text view of an interface and act on screenshots through a vision-language screen reader. Yet on a simulated mobile platform with a low-cost reader, policies that complete most tasks in text fail on many of the same tasks on screenshots. We ask where a small budget of real screenshots is best spent to close this gap. We find that the reader's misreads are systematic and recur on the same interface elements. Noise matched to the reader's error rates, a natural sim-to-real remedy, misses this structure, and in a pre-specified comparison it did not beat training the policy directly on screenshots. We therefore correct the misreads at their source with reader alignment, which fine-tunes the reader on screenshots paired with their text views. The reader learns to output each screen's text view, as in text practice, and the policy stays unchanged. Spent on the reader, the screenshot budget gives 80.9% full-task success on new task instances, against 55.6% when spent on training the policy. Given the same labeled screenshots, adapting the reader also beats adapting the policy, and a single aligned reader improves every text-trained policy we tried without retraining any of them.

Community

Sign up or log in to comment

Models citing this paper 1

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2611.22722 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.22722 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.