On the Effects of Data Scale on UI Control Agents

TL;DR

This paper investigates how increasing training data affects the performance of fine-tuned language models controlling Android apps, revealing that in-domain performance improves with more data, but out-of-domain performance remains challenging.

Contribution

The study introduces AndroidControl, a large diverse dataset for training and analyzing UI control agents, and provides insights into data scale effects on in-domain and out-of-domain performance.

Findings

In-domain fine-tuned models improve with more data.

Out-of-domain performance improves slowly and remains limited.

More data alone may not suffice for robust out-of-domain control.

Abstract

Autonomous agents that control computer interfaces to accomplish human tasks are emerging. Leveraging LLMs to power such agents has been of special interest, but unless fine-tuned on human-collected task demonstrations, performance is still relatively low. In this work we study whether fine-tuning alone is a viable approach for building real-world computer control agents. In particularly, we investigate how performance measured on both high and low-level tasks in domain and out of domain scales as more training data is collected. To this end we collect and release a new dataset, AndroidControl, consisting of 15,283 demonstrations of everyday tasks with Android apps. Compared to existing datasets, each AndroidControl task instance includes both high and low-level human-generated instructions, allowing us to explore the level of task complexity an agent can handle. Moreover, AndroidControl is the most diverse computer control dataset to date, including 14,548 unique tasks over 833 Android apps, thus allowing us to conduct in-depth analysis of the model performance in and out of the domain of the training data. Using the dataset, we find that when tested in domain fine-tuned models outperform zero and few-shot baselines and scale in such a way that robust performance might feasibly be obtained simply by collecting more data. Out of domain, performance scales significantly more slowly and suggests that in particular for high-level tasks, fine-tuning on more data alone may be insufficient for achieving robust out-of-domain performance.