Never-ending Learning of User Interfaces

TL;DR

The paper introduces the Never-ending UI Learner, an automated system that crawls real apps to generate training data for predicting UI element properties, reducing reliance on human-labeled datasets.

Contribution

It presents a novel automated app crawling approach that continuously collects training data for UI understanding models, improving data quality and coverage.

Findings

Crawled over 6,000 apps for 5,000+ device-hours.

Trained models for tappability, draggability, and screen similarity.

Performed over 500,000 actions to gather diverse UI data.

Abstract

Machine learning models have been trained to predict semantic information about user interfaces (UIs) to make apps more accessible, easier to test, and to automate. Currently, most models rely on datasets that are collected and labeled by human crowd-workers, a process that is costly and surprisingly error-prone for certain tasks. For example, it is possible to guess if a UI element is "tappable" from a screenshot (i.e., based on visual signifiers) or from potentially unreliable metadata (e.g., a view hierarchy), but one way to know for certain is to programmatically tap the UI element and observe the effects. We built the Never-ending UI Learner, an app crawler that automatically installs real apps from a mobile app store and crawls them to discover new and challenging training examples to learn from. The Never-ending UI Learner has crawled for more than 5,000 device-hours, performing over half a million actions on 6,000 apps to train three computer vision models for i) tappability prediction, ii) draggability prediction, and iii) screen similarity.