MobileFlow: A Multimodal LLM For Mobile GUI Agent

TL;DR

MobileFlow is a multimodal large language model designed for mobile GUI agents, supporting variable image resolutions and multilingual interfaces, outperforming existing models in GUI task execution and practical deployment.

Contribution

The paper introduces MobileFlow, a novel 21-billion-parameter multimodal LLM with hybrid visual encoders and innovative training strategies tailored for mobile GUI understanding and interaction, especially in Chinese.

Findings

Outperforms Qwen-VL-Max and GPT-4v in GUI task execution.

Supports variable image resolutions and multilingual GUIs.

Successfully deployed in real-world business applications.

Abstract

Currently, the integration of mobile Graphical User Interfaces (GUIs) is ubiquitous in most people's daily lives. And the ongoing evolution of multimodal large-scale models, such as GPT-4v, Qwen-VL-Max, has significantly bolstered the capabilities of GUI comprehension and user action analysis, showcasing the potentiality of intelligent GUI assistants. However, current GUI Agents often need to access page layout information through calling system APIs, which may pose privacy risks. Fixing GUI (such as mobile interfaces) to a certain low resolution might result in the loss of fine-grained image details. At the same time, the multimodal large models built for GUI Agents currently have poor understanding and decision-making abilities for Chinese GUI interfaces, making them difficult to apply to a large number of Chinese apps. This paper introduces MobileFlow, a multimodal large language model meticulously crafted for mobile GUI agents. Transforming from the open-source model Qwen-VL-Chat into GUI domain, MobileFlow contains approximately 21 billion parameters and is equipped with novel hybrid visual encoders, making it possible for variable resolutions of image inputs and good support for multilingual GUI. By incorporating Mixture of Experts (MoE) expansions and pioneering alignment training strategies, MobileFlow has the capacity to fully interpret image data and comprehend user instructions for GUI interaction tasks. Finally, MobileFlow outperforms Qwen-VL-Max and GPT-4v in terms of task execution by GUI agents on both public and our proposed evaluation metrics, and has been successfully deployed in real-world business contexts, proving its effectiveness for practical applications.