BrowserAgent: Building Web Agents with Human-Inspired Web Browsing Actions
Published in Under Review, 2025
Recommended citation: Yu, T., Zhang, Z., Lyu, Z., Gong, J., Yi, H., Wang, X., Zhou, Y., Yang, J., Nie, P., Huang, Y., & Chen, W. (2025). BrowserAgent: Building Web Agents with Human-Inspired Web Browsing Actions. Under Review. https://arxiv.org/abs/placeholder
BrowserAgent focuses on information-seeking tasks on the web browser through direct environment interaction, unlike traditional DeepResearch-style agents that rely on static tool outputs. This work represents a significant advancement in agentic post-training for web-based tasks.
The framework enables agents to perform complex web navigation and information extraction tasks by learning human-inspired browsing behaviors and interactions.
Recommended citation:
@article{yu2025browseragent,
title={BrowserAgent: Building Web Agents with Human-Inspired Web Browsing Actions},
author={Yu, Tao and Zhang, Zhengbo and Lyu, Zhiheng and Gong, Jialei and Yi, Hang and Wang, Xin and Zhou, Yiran and Yang, Jingwen and Nie, Pengyu and Huang, Yilun and Chen, Wenhu},
journal={arXiv preprint},
year={2025}
}