ScreenAgent: A Vision Language Model-driven Computer Control Agent

TL;DR

This paper introduces ScreenAgent, a vision language model-driven agent capable of controlling a computer by interacting with GUI elements through screenshots and actions, aiming to develop a generalized digital assistant.

Contribution

It constructs a new environment and dataset for training vision language models to control computers, and demonstrates a model with capabilities comparable to GPT-4V in UI interaction.

Findings

ScreenAgent achieves computer control comparable to GPT-4V.

The model demonstrates precise UI positioning capabilities.

The environment and dataset facilitate future research in generalist LLM agents.

Abstract

Existing Large Language Models (LLM) can invoke a variety of tools and APIs to complete complex tasks. The computer, as the most powerful and universal tool, could potentially be controlled directly by a trained LLM agent. Powered by the computer, we can hopefully build a more generalized agent to assist humans in various daily digital works. In this paper, we construct an environment for a Vision Language Model (VLM) agent to interact with a real computer screen. Within this environment, the agent can observe screenshots and manipulate the Graphics User Interface (GUI) by outputting mouse and keyboard actions. We also design an automated control pipeline that includes planning, acting, and reflecting phases, guiding the agent to continuously interact with the environment and complete multi-step tasks. Additionally, we construct the ScreenAgent Dataset, which collects screenshots and action sequences when completing a variety of daily computer tasks. Finally, we trained a model, ScreenAgent, which achieved computer control capabilities comparable to GPT-4V and demonstrated more precise UI positioning capabilities. Our attempts could inspire further research on building a generalist LLM agent. The code is available at \url{https://github.com/niuzaisheng/ScreenAgent}.