---
title: "ColorBrowserAgent — paper by Wang et al. (2026)"
description: "The authors, from OPPO Research Institute and Shanghai Jiao Tong University, propose ColorBrowserAgent, which adapts… 42 references."
authors: ["Jihong Wang", "Jiamu Zhou", "Weiming Zhang", "Teng Wang", "Weiwen Liu", "Zhuosheng Zhang", "Xingyu Lou", "Weinan Zhang", "Huarong Deng", "Jun Wang"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track)"
source: "https://aclanthology.org/2026.acl-industry.46/"
doi: "10.18653/v1/2026.acl-industry.46"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/wang-2026-colorbrowseragent-complex-long-horizon-browser/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/wang-2026-colorbrowseragent-complex-long-horizon-browser/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/11/) / ColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution

# ColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution

**Authors:** Jihong Wang, Jiamu Zhou, [Weiming Zhang](https://md.research.elamur.ai/free/en/authors/weiming-zhang/), [Teng Wang](https://md.research.elamur.ai/free/en/authors/teng-wang/), [Weiwen Liu](https://md.research.elamur.ai/free/en/authors/weiwen-liu/), [Zhuosheng Zhang](https://md.research.elamur.ai/free/en/authors/zhuosheng-zhang/), Xingyu Lou, [Weinan Zhang](https://md.research.elamur.ai/free/en/authors/weinan-zhang/), Huarong Deng and [Jun Wang](https://md.research.elamur.ai/free/en/authors/jun-wang/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-industry.46/) · [PDF](https://aclanthology.org/2026.acl-industry.46.pdf) · [DOI 10.18653/v1/2026.acl-industry.46](https://doi.org/10.18653/v1/2026.acl-industry.46) · [OpenAlex](https://openalex.org/W7166869424)

**Citations:** 1 citations (OpenAlex, 2026-10-14)

## Summary
The paper addresses two obstacles to deploying web-browsing agents in production. First, general-purpose models lack prior knowledge of the many different site layouts and business rules. Second, agents tend to drift away from the goal as interaction sequences grow longer. The authors, from OPPO Research Institute and Shanghai Jiao Tong University, propose ColorBrowserAgent, which adapts without fine-tuning.

In an offline loop, an Adaptor detects failures and passes them to human experts. The experts write short, reusable tips about each site's operating logic, which are stored in an Adaptive Knowledge Base and retrieved at run time. In the online loop, a Summarizer keeps a compressed belief state and checks planned actions against the retrieved tips, and an Operator then executes the browser actions.

The system reaches state-of-the-art results on WebArena and transfers its frozen knowledge zero-shot to the harder WebChoreArena. In a commercial travel-planning service it noticeably raised user satisfaction. This suggests that storing domain knowledge outside model weights is a cheap, maintainable alternative to retraining.

## Key points
- **Benchmark results:** 71.2% success on WebArena, the best reported, though Map lags at 55.9% versus 64.2% for the top baseline. On WebChoreArena, zero-shot transfer gives 47.4%, compared with 31.1% for WebDART and 34.4% for a knowledge-free variant, all using GPT-5-based setups.
- **Ablation:** On WebArena-Lite (165 tasks, GPT-5), the full system scores 72.6%. Removing the Summarizer drops it to 68.8%, removing the Adaptor drops it to 65.4%, and the vanilla agent scores 61.7%.
- **Knowledge retrieval:** The Adaptor retrieves tips through a cascade of URL pattern matching, then keyword search, then visual-semantic embedding. Failures are flagged by a hybrid trigger that combines rules with a VLM-based evaluator.
- **Memory compression:** The Summarizer keeps full detail only for the active subgoal and collapses finished steps into summaries. This keeps memory roughly constant in size.
- **Deployment:** In an A/B test against a search-engine pipeline, user satisfaction rose 19.3% relative to the baseline, and task success exceeded 95% in verticals with accumulated priors. Stated limitations are cold-start effort on new domains and dependence on proprietary VLMs such as GPT-5.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Shanghai Jiao Tong University](https://md.research.elamur.ai/free/en/universities/shanghai-jiao-tong-university/) — University, Shanghai, China · 149 works · 16,161 citations

📚 Shelves (2): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,437 works

## Cite

Cite as: Jihong Wang, Jiamu Zhou, Weiming Zhang, Teng Wang, Weiwen Liu, Zhuosheng Zhang, Xingyu Lou, Weinan Zhang, Huarong Deng and Jun Wang. 2026. ColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{wang2026colorbrowseragent,
  title = {{ColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution}},
  author = {Jihong Wang, Jiamu Zhou, Weiming Zhang, Teng Wang, Weiwen Liu, Zhuosheng Zhang, Xingyu Lou, Weinan Zhang, Huarong Deng and Jun Wang},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-industry.46},
  url = {https://doi.org/10.18653/v1/2026.acl-industry.46}
}
```

- [← previous Cohesion-6K: An Arabic Dataset for Analyzing Social Cohesion and Conflict in Online Discourse](https://md.research.elamur.ai/free/en/papers/athba-2026-cohesion-6k-arabic-dataset-analyzing/)
- [next → ComfyUI-R1: Exploring Reasoning Models for Workflow Generation](https://md.research.elamur.ai/free/en/papers/xu-2026-comfyui-r1-exploring-reasoning-models/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/11/)

## References

42 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, and 1 others. 2023. [Gpt-4 technical report](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/). arXiv preprint arXiv:2303.08774.
2. De Chezelles, Thibault Le Sellier, Sahar Omidi Shayegan, Lawrence Keunho Jang, Xing Han Lu, Ori Yoran, Dehan Kong, Frank F Xu, Siva Reddy, Quentin Cappart, and 1 others. 2024. The browsergym ecosystem for web agent research. arXiv preprint arXiv:2412.05467. [find](https://scholar.google.com/scholar?q=De+Chezelles%2C+Thibault+Le+Sellier%2C+Sahar+Omidi+Shayegan%2C+Lawrence+Keunho+Jang%2C+Xing+Han+Lu%2C+Ori+Yoran%2C+Dehan+Kong%2C+Frank+F+Xu%2C+Siva+Reddy%2C+Quentin+Cappart%2C+and+1+others%2E+2024%2E+The+browsergym+ecosystem)
3. Thibault Le Sellier De Chezelles, Maxime Gasse, Alexandre Drouin, Massimo Caccia, Léo Boisvert, Megh Thakkar, Tom Marty, Rim Assouel, Sahar Omidi Shayegan, Lawrence Keunho Jang, Xing Han Lù, Ori Yoran, Dehan Kong, Frank F. Xu, Siva Reddy, Quentin Cappart, Graham Neubig, Ruslan Salakhutdinov, Nicolas… [find](https://scholar.google.com/scholar?q=Thibault+Le+Sellier+De+Chezelles%2C+Maxime+Gasse%2C+Alexandre+Drouin%2C+Massimo+Caccia%2C+L%C3%A9o+Boisvert%2C+Megh+Thakkar%2C+Tom+Marty%2C+Rim+Assouel%2C+Sahar+Omidi+Shayegan%2C+Lawrence+Keunho+Jang%2C+Xing+Han+L%C3%B9%2C+Ori+Yoran)
4. Mahir Labib Dihan, Tanzima Hashem, Mohammed Eunus Ali, and Md Rizwan Parvez. 2025. Weboperator: Action-aware tree search for autonomous agents in web environment. arXiv preprint arXiv:2512.12692. — [arXiv:2512.12692](https://arxiv.org/abs/2512.12692) [find](https://scholar.google.com/scholar?q=Mahir+Labib+Dihan%2C+Tanzima+Hashem%2C+Mohammed+Eunus+Ali%2C+and+Md+Rizwan+Parvez%2E+2025%2E+Weboperator%3A+Action%2Daware+tree+search+for+autonomous+agents+in+web+environment%2E+arXiv+preprint+arXiv%3A2512%2E12692%2E+%E2%80%94+%5Ba)
5. A. Drouin, M. Gasse, M. Caccia, and 1 others. 2024. Workarena: How capable are web agents at solving common knowledge work tasks? In Forty-first International Conference on Machine Learning. [find](https://scholar.google.com/scholar?q=A%2E+Drouin%2C+M%2E+Gasse%2C+M%2E+Caccia%2C+and+1+others%2E+2024%2E+Workarena%3A+How+capable+are+web+agents+at+solving+common+knowledge+work+tasks%3F+In+Forty%2Dfirst+International+Conference+on+Machine+Learning%2E)
6. A. Gandhi, G. Neubig, and M. L. Dihan. 2025. Go-browse: Training web agents with structured exploration. arXiv preprint arXiv:2506.03533. [find](https://scholar.google.com/scholar?q=A%2E+Gandhi%2C+G%2E+Neubig%2C+and+M%2E+L%2E+Dihan%2E+2025%2E+Go%2Dbrowse%3A+Training+web+agents+with+structured+exploration%2E+arXiv+preprint+arXiv%3A2506%2E03533%2E)
7. S. He, Y. Cui, X. Ma, and 1 others. 2025. Branch-and-browse: Efficient and controllable web exploration with tree-structured reasoning and action memory. arXiv preprint arXiv:2510.19838. [find](https://scholar.google.com/scholar?q=S%2E+He%2C+Y%2E+Cui%2C+X%2E+Ma%2C+and+1+others%2E+2025%2E+Branch%2Dand%2Dbrowse%3A+Efficient+and+controllable+web+exploration+with+tree%2Dstructured+reasoning+and+action+memory%2E+arXiv+preprint+arXiv%3A2510%2E19838%2E)
8. Su Kara, Fazle Faisal, and Suman Nath. 2025. Waber: Evaluating reliability and efficiency of web agents with existing benchmarks. In ICLR 2025 Workshop on Foundation Models in the Wild. [find](https://scholar.google.com/scholar?q=Su+Kara%2C+Fazle+Faisal%2C+and+Suman+Nath%2E+2025%2E+Waber%3A+Evaluating+reliability+and+efficiency+of+web+agents+with+existing+benchmarks%2E+In+ICLR+2025+Workshop+on+Foundation+Models+in+the+Wild%2E)
9. J. Y. Koh, S. McAleer, D. Fried, and 1 others. 2024. Tree search for language model agents. arXiv preprint arXiv:2407.01476. [find](https://scholar.google.com/scholar?q=J%2E+Y%2E+Koh%2C+S%2E+McAleer%2C+D%2E+Fried%2C+and+1+others%2E+2024%2E+Tree+search+for+language+model+agents%2E+arXiv+preprint+arXiv%3A2407%2E01476%2E)
10. Ning Li, Qiqiang Lin, Zheng Wu, Xiaoyun Mo, Weiming Zhang, Yin Zhao, Xiangmou Qu, Jiamu Zhou, Jun Wang, Congmin Zheng, and 1 others. 2025. Col-oragent: Building a robust, personalized, and interactive os agent. arXiv preprint arXiv:2510.19386. — [arXiv:2510.19386](https://arxiv.org/abs/2510.19386) [find](https://scholar.google.com/scholar?q=Ning+Li%2C+Qiqiang+Lin%2C+Zheng+Wu%2C+Xiaoyun+Mo%2C+Weiming+Zhang%2C+Yin+Zhao%2C+Xiangmou+Qu%2C+Jiamu+Zhou%2C+Jun+Wang%2C+Congmin+Zheng%2C+and+1+others%2E+2025%2E+Col%2Doragent%3A+Building+a+robust%2C+personalized%2C+and+interactive)
11. Yanda Li, Chi Zhang, Wenjia Jiang, Wanqi Yang, Bin Fu, Pei Cheng, Xin Chen, Ling Chen, and Yunchao Wei. 2024. Appagent v2: Advanced agent for flexible mobile interactions. arXiv preprint arXiv:2408.11824. — [arXiv:2408.11824](https://arxiv.org/abs/2408.11824) [find](https://scholar.google.com/scholar?q=Yanda+Li%2C+Chi+Zhang%2C+Wenjia+Jiang%2C+Wanqi+Yang%2C+Bin+Fu%2C+Pei+Cheng%2C+Xin+Chen%2C+Ling+Chen%2C+and+Yunchao+Wei%2E+2024%2E+Appagent+v2%3A+Advanced+agent+for+flexible+mobile+interactions%2E+arXiv+preprint+arXiv%3A2408%2E11)
12. Junteng Liu, Yunji Li, Chi Zhang, Jingyang Li, Aili Chen, Ke Ji, Weiyu Cheng, Zijia Wu, Chengyu Du, Qidi Xu, Jiayuan Song, Zhengmao Zhu, Wenhu Chen, Pengyu Zhao, and Junxian He. 2025a. Webexplorer: Explore and evolve for training long-horizon web agents. arXiv preprint arXiv:2509.06501. — [arXiv:2509.06501](https://arxiv.org/abs/2509.06501) [find](https://scholar.google.com/scholar?q=Junteng+Liu%2C+Yunji+Li%2C+Chi+Zhang%2C+Jingyang+Li%2C+Aili+Chen%2C+Ke+Ji%2C+Weiyu+Cheng%2C+Zijia+Wu%2C+Chengyu+Du%2C+Qidi+Xu%2C+Jiayuan+Song%2C+Zhengmao+Zhu%2C+Wenhu+Chen%2C+Pengyu+Zhao%2C+and+Junxian+He%2E+2025a%2E+Webexplorer%3A+Ex)
13. Xiao Liu and 1 others. 2025b. Webcoach: Learning from human feedback via cross-session memory guidance. Nature Machine Intelligence (In Press). [find](https://scholar.google.com/scholar?q=Xiao+Liu+and+1+others%2E+2025b%2E+Webcoach%3A+Learning+from+human+feedback+via+cross%2Dsession+memory+guidance%2E+Nature+Machine+Intelligence+%28In+Press%29%2E)
14. SHengjie Ma, Chenlong Deng, Jiaxin Mao, Jiadeng Huang, Teng Wang, Junjie Wu, Changwang Zhang, and 1 others. 2025. Pou: Proof-of-use to counter tool-call hacking in deepresearch agents. arXiv preprint arXiv:2510.10931. — [arXiv:2510.10931](https://arxiv.org/abs/2510.10931) [find](https://scholar.google.com/scholar?q=SHengjie+Ma%2C+Chenlong+Deng%2C+Jiaxin+Mao%2C+Jiadeng+Huang%2C+Teng+Wang%2C+Junjie+Wu%2C+Changwang+Zhang%2C+and+1+others%2E+2025%2E+Pou%3A+Proof%2Dof%2Duse+to+counter+tool%2Dcall+hacking+in+deepresearch+agents%2E+arXiv+preprint+)
15. Atsuyuki Miyai, Zaiying Zhao, Kazuki Egashira, Atsuki Sato, Tatsumi Sunada, Shota Onohara, Hiromasa Yamanishi, Mashiro Toyooka, Kunato Nishina, Ryoma Maeda, and 1 others. 2025. Webchorearena: Evaluating web browsing agents on realistic tedious web tasks. arXiv preprint arXiv:2506.01952. — [arXiv:2506.01952](https://arxiv.org/abs/2506.01952) [find](https://scholar.google.com/scholar?q=Atsuyuki+Miyai%2C+Zaiying+Zhao%2C+Kazuki+Egashira%2C+Atsuki+Sato%2C+Tatsumi+Sunada%2C+Shota+Onohara%2C+Hiromasa+Yamanishi%2C+Mashiro+Toyooka%2C+Kunato+Nishina%2C+Ryoma+Maeda%2C+and+1+others%2E+2025%2E+Webchorearena%3A+Evaluati)
16. Liangbo Ning, Ziran Liang, Zhuohang Jiang, Haohao Qu, Yujuan Ding, Wenqi Fan, Xiao-yong Wei, Shanru Lin, Hui Liu, Philip S Yu, and 1 others. 2025. A survey of webagents: Towards next-generation ai agents for web automation with large foundation models. In Proceedings of the 31st ACM SIGKDD Conferenc… [find](https://scholar.google.com/scholar?q=Liangbo+Ning%2C+Ziran+Liang%2C+Zhuohang+Jiang%2C+Haohao+Qu%2C+Yujuan+Ding%2C+Wenqi+Fan%2C+Xiao%2Dyong+Wei%2C+Shanru+Lin%2C+Hui+Liu%2C+Philip+S+Yu%2C+and+1+others%2E+2025%2E+A+survey+of+webagents%3A+Towards+next%2Dgeneration+ai+age)
17. OpenAI. 2025. Gpt-5 is here. [https://openai.com/](https://openai.com/) gpt-5/. [↗](https://openai.com/)
18. Zehan Qi, Xiao Liu, Iat Long Iong, Hanyu Lai, Xueqiao Sun, Wenyi Zhao, Yu Yang, Xinyue Yang, Jiadai Sun, Shuntian Yao, and 1 others. 2024. Webrl: Training llm web agents via self-evolving online curriculum reinforcement learning. arXiv preprint arXiv:2411.02337. — [arXiv:2411.02337](https://arxiv.org/abs/2411.02337) [find](https://scholar.google.com/scholar?q=Zehan+Qi%2C+Xiao+Liu%2C+Iat+Long+Iong%2C+Hanyu+Lai%2C+Xueqiao+Sun%2C+Wenyi+Zhao%2C+Yu+Yang%2C+Xinyue+Yang%2C+Jiadai+Sun%2C+Shuntian+Yao%2C+and+1+others%2E+2024%2E+Webrl%3A+Training+llm+web+agents+via+self%2Devolving+online+curri)
19. J. Shen, A. Jain, Z. Xiao, and 1 others. 2024. Scribeagent: Towards specialized web agents using production-scale workflow data. arXiv preprint arXiv:2411.15004. — [arXiv:2411.15004](https://arxiv.org/abs/2411.15004) [find](https://scholar.google.com/scholar?q=J%2E+Shen%2C+A%2E+Jain%2C+Z%2E+Xiao%2C+and+1+others%2E+2024%2E+Scribeagent%3A+Towards+specialized+web+agents+using+production%2Dscale+workflow+data%2E+arXiv+preprint+arXiv%3A2411%2E15004%2E+%E2%80%94+%5BarXiv%3A2411%2E15004%5D%28https%3A%2F%2Farxiv%2Eorg)
20. Segev Shlomov, Alon Oved, Sami Marreed, Ido Levy, Offer Akrabi, Avi Yaeli, Łukasz Str ak, Elizabeth Koumpan, Yinon Goldshtein, Eilam Shapira, and 1 others. 2025. From benchmarks to business impact: Deploying ibm generalist agent in enterprise production. arXiv preprint arXiv:2510.23856. — [arXiv:2510.23856](https://arxiv.org/abs/2510.23856) [find](https://scholar.google.com/scholar?q=Segev+Shlomov%2C+Alon+Oved%2C+Sami+Marreed%2C+Ido+Levy%2C+Offer+Akrabi%2C+Avi+Yaeli%2C+%C5%81ukasz+Str+ak%2C+Elizabeth+Koumpan%2C+Yinon+Goldshtein%2C+Eilam+Shapira%2C+and+1+others%2E+2025%2E+From+benchmarks+to+business+impact%3A+De)
21. Paloma Sodhi, SRK Branavan, Yoav Artzi, and Ryan McDonald. 2023. Step: Stacked llm policies for web actions. arXiv preprint arXiv:2310.03720. — [arXiv:2310.03720](https://arxiv.org/abs/2310.03720) [find](https://scholar.google.com/scholar?q=Paloma+Sodhi%2C+SRK+Branavan%2C+Yoav+Artzi%2C+and+Ryan+McDonald%2E+2023%2E+Step%3A+Stacked+llm+policies+for+web+actions%2E+arXiv+preprint+arXiv%3A2310%2E03720%2E+%E2%80%94+%5BarXiv%3A2310%2E03720%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2310%2E03720%29)
22. Haoming Wang, Haoyang Zou, Huatong Song, Jiazhan Feng, Junjie Fang, Junting Lu, Longxiang Liu, Qinyu Luo, Shihao Liang, Shijue Huang, and 1 others. 2025a. [Ui-tars-2 technical report: Advancing gui agent with multi-turn reinforcement learning](https://md.research.elamur.ai/free/en/papers/wang-2025-ui-tars-2-technical/). arXiv preprint arXiv:2509.02544.
23. Teng Wang, Zhangyi Jiang, Zhenqi He, Shenyang Tong, Wenhan Yang, Yanan Zheng, Zeyu Li, Zifan He, Hailei Gong, Zewen Ye, and 1 others. 2025b. [Towards hierarchical multi-step reward models for enhanced reasoning in large language models](https://md.research.elamur.ai/free/en/papers/wang-2026-hierarchical-multi-step-reward-models/). arXiv preprint arXiv:2503.13551.
24. Teng Wang, Rong Shan, Jianghao Lin, Junjie Wu, Tianyi Xu, Jianping Zhang, Wenteng Chen, Changwang Zhang, Zhaoxiang Wang, Weinan Zhang, and Jun Wang. 2026. Oscar: Optimization-steered agentic planning for composed image retrieval. Preprint, arXiv:2602.08603. 2025c. Bpp-search: Enhancing tree of thoug… [find](https://scholar.google.com/scholar?q=Teng+Wang%2C+Rong+Shan%2C+Jianghao+Lin%2C+Junjie+Wu%2C+Tianyi+Xu%2C+Jianping+Zhang%2C+Wenteng+Chen%2C+Changwang+Zhang%2C+Zhaoxiang+Wang%2C+Weinan+Zhang%2C+and+Jun+Wang%2E+2026%2E+Oscar%3A+Optimization%2Dsteered+agentic+planning+)
25. Z. Z. Wang, J. Y. Mao, D. Fried, and 1 others. 2024a. [Agent workflow memory](https://md.research.elamur.ai/free/en/papers/wang-2024-agent-workflow-memory/). arXiv preprint arXiv:2409.07429.
26. Zora Zhiruo Wang, Jiayuan Mao, Daniel Fried, and Graham Neubig. 2024b. [Agent workflow memory](https://md.research.elamur.ai/free/en/papers/wang-2024-agent-workflow-memory/). arXiv preprint arXiv:2409.07429.
27. Zhepei Wei, Wenlin Yao, Yao Liu, Weizhi Zhang, Qin Lu, Liang Qiu, Changlong Yu, Puyang Xu, Chao Zhang, Bing Yin, and 1 others. 2025. [Webagent-r1: Training web agents via end-to-end multi-turn reinforcement learning](https://md.research.elamur.ai/free/en/papers/wei-2025-webagent-r1-training-web-agents/). arXiv preprint arXiv:2505.16421.
28. Xingjiao Wu, Luwei Xiao, Yixuan Sun, Junhang Zhang, Tianlong Ma, and Liang He. 2022. A survey of human-in-the-loop for machine learning. Future Generation Computer Systems, 135:364–381. [find](https://scholar.google.com/scholar?q=Xingjiao+Wu%2C+Luwei+Xiao%2C+Yixuan+Sun%2C+Junhang+Zhang%2C+Tianlong+Ma%2C+and+Liang+He%2E+2022%2E+A+survey+of+human%2Din%2Dthe%2Dloop+for+machine+learning%2E+Future+Generation+Computer+Systems%2C+135%3A364%E2%80%93381%2E)
29. Tianci Xue, Weijian Qi, Tianneng Shi, Chan Hee Song, Boyu Gou, Dawn Song, Huan Sun, and Yu Su. 2025. An illusion of progress? assessing the current state of web agents. arXiv preprint arXiv:2504.01382. [find](https://scholar.google.com/scholar?q=Tianci+Xue%2C+Weijian+Qi%2C+Tianneng+Shi%2C+Chan+Hee+Song%2C+Boyu+Gou%2C+Dawn+Song%2C+Huan+Sun%2C+and+Yu+Su%2E+2025%2E+An+illusion+of+progress%3F+assessing+the+current+state+of+web+agents%2E+arXiv+preprint+arXiv%3A2504%2E01382)
30. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, and 1 others. 2025a. [Qwen3 technical report](https://md.research.elamur.ai/free/en/papers/yang-2025-qwen3-technical-report/). arXiv preprint arXiv:2505.09388.
31. Jianwei Yang, Hao Zhang, Feng Li, Xueyan Zou, Chunyuan Li, and Jianfeng Gao. 2023. Set-of-mark prompting unleashes extraordinary visual grounding in gpt-4v. arXiv preprint arXiv:2310.11441. [find](https://scholar.google.com/scholar?q=Jianwei+Yang%2C+Hao+Zhang%2C+Feng+Li%2C+Xueyan+Zou%2C+Chunyuan+Li%2C+and+Jianfeng+Gao%2E+2023%2E+Set%2Dof%2Dmark+prompting+unleashes+extraordinary+visual+grounding+in+gpt%2D4v%2E+arXiv+preprint+arXiv%3A2310%2E11441%2E)
32. Jingbo Yang, Bairu Hou, Wei Wei, Shiyu Chang, and Yujia Bao. 2025b. [Webdart: Dynamic decomposition and re-planning for complex web tasks](https://md.research.elamur.ai/free/en/papers/yang-2026-webdart-dynamic-decomposition-re-planning/). arXiv preprint arXiv:2510.06587.
33. K. Yang, Y. Liu, S. Chaudhary, and 1 others. 2025c. Agentoccam: A simple yet strong baseline for llm-based web agents. In The Thirteenth International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=K%2E+Yang%2C+Y%2E+Liu%2C+S%2E+Chaudhary%2C+and+1+others%2E+2025c%2E+Agentoccam%3A+A+simple+yet+strong+baseline+for+llm%2Dbased+web+agents%2E+In+The+Thirteenth+International+Conference+on+Learning+Representations%2E)
34. Ke Yang, Yao Liu, Sapana Chaudhary, Rasool Fakoor, Pratik Chaudhari, George Karypis, and Huzefa Rangwala. 2024. Agentoccam: A simple yet strong baseline for llm-based web agents. arXiv preprint arXiv:2410.13825. — [arXiv:2410.13825](https://arxiv.org/abs/2410.13825) [find](https://scholar.google.com/scholar?q=Ke+Yang%2C+Yao+Liu%2C+Sapana+Chaudhary%2C+Rasool+Fakoor%2C+Pratik+Chaudhari%2C+George+Karypis%2C+and+Huzefa+Rangwala%2E+2024%2E+Agentoccam%3A+A+simple+yet+strong+baseline+for+llm%2Dbased+web+agents%2E+arXiv+preprint+arXiv%3A)
35. Fabio Massimo Zanzotto. 2019. Human-in-the-loop artificial intelligence. Journal of Artificial Intelligence Research, 64:243–252. [find](https://scholar.google.com/scholar?q=Fabio+Massimo+Zanzotto%2E+2019%2E+Human%2Din%2Dthe%2Dloop+artificial+intelligence%2E+Journal+of+Artificial+Intelligence+Research%2C+64%3A243%E2%80%93252%2E)
36. Teng Wang, Wing-Yin Yu, Zhenqi He, Zehua Liu, HaileiGong HaileiGong, Han Wu, Xiongwei Han, Wei Shi, Ruifeng She, Fangzhou Zhu, and 1 others. [find](https://scholar.google.com/scholar?q=Teng+Wang%2C+Wing%2DYin+Yu%2C+Zhenqi+He%2C+Zehua+Liu%2C+HaileiGong+HaileiGong%2C+Han+Wu%2C+Xiongwei+Han%2C+Wei+Shi%2C+Ruifeng+She%2C+Fangzhou+Zhu%2C+and+1+others%2E)
37. Chaoyun Zhang, Shilin He, Jiaxu Qian, Bowen Li, Liqun Li, Si Qin, Yu Kang, Minghua Ma, Guyue Liu, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, and Qi Zhang. 2024. Large language model-brained gui agents: A survey. arXiv preprint arXiv:2411.18279. — [arXiv:2411.18279](https://arxiv.org/abs/2411.18279) [find](https://scholar.google.com/scholar?q=Chaoyun+Zhang%2C+Shilin+He%2C+Jiaxu+Qian%2C+Bowen+Li%2C+Liqun+Li%2C+Si+Qin%2C+Yu+Kang%2C+Minghua+Ma%2C+Guyue+Liu%2C+Qingwei+Lin%2C+Saravan+Rajmohan%2C+Dongmei+Zhang%2C+and+Qi+Zhang%2E+2024%2E+Large+language+model%2Dbrained+gui+age)
38. Chi Zhang, Zhao Yang, Jiaxuan Liu, Yanda Li, Yucheng Han, Xin Chen, Zebiao Huang, Bin Fu, and Gang Yu. 2025a. Appagent: Multimodal agents as smartphone users. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, pages 1– 20. [find](https://scholar.google.com/scholar?q=Chi+Zhang%2C+Zhao+Yang%2C+Jiaxuan+Liu%2C+Yanda+Li%2C+Yucheng+Han%2C+Xin+Chen%2C+Zebiao+Huang%2C+Bin+Fu%2C+and+Gang+Yu%2E+2025a%2E+Appagent%3A+Multimodal+agents+as+smartphone+users%2E+In+Proceedings+of+the+2025+CHI+Conference)
39. R. Zhang, M. Qiu, Z. Tan, and 1 others. 2025b. Symbiotic cooperation for web agents: Harnessing complementary strengths of large and small llms. CoRR. [find](https://scholar.google.com/scholar?q=R%2E+Zhang%2C+M%2E+Qiu%2C+Z%2E+Tan%2C+and+1+others%2E+2025b%2E+Symbiotic+cooperation+for+web+agents%3A+Harnessing+complementary+strengths+of+large+and+small+llms%2E+CoRR%2E)
40. Y. Zhang, Z. Ma, Y. Ma, and 1 others. 2025c. Webpilot: A versatile and autonomous multi-agent system for web task execution with strategic exploration. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 23378–23386. [find](https://scholar.google.com/scholar?q=Y%2E+Zhang%2C+Z%2E+Ma%2C+Y%2E+Ma%2C+and+1+others%2E+2025c%2E+Webpilot%3A+A+versatile+and+autonomous+multi%2Dagent+system+for+web+task+execution+with+strategic+exploration%2E+In+Proceedings+of+the+AAAI+Conference+on+Artific)
41. Yao Zhang, Zijian Ma, Yunpu Ma, Zhen Han, Yu Wu, and Volker Tresp. 2025d. Webpilot: A versatile and autonomous multi-agent system for web task execution with strategic exploration. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 23378–23386. [find](https://scholar.google.com/scholar?q=Yao+Zhang%2C+Zijian+Ma%2C+Yunpu+Ma%2C+Zhen+Han%2C+Yu+Wu%2C+and+Volker+Tresp%2E+2025d%2E+Webpilot%3A+A+versatile+and+autonomous+multi%2Dagent+system+for+web+task+execution+with+strategic+exploration%2E+In+Proceedings+of+t)
42. Shuyan Zhou, Frank F. Xu, Hao Zhu, Xuhui Zhou, Robert Lo, Abishek Sridhar, and Graham Neubig. 2023. [Webarena: A realistic web environment for building autonomous agents](https://md.research.elamur.ai/free/en/papers/zhou-2023-webarena-realistic-web-environment-building/). arXiv preprint arXiv:2307.13854.

*Updated 1 November 2026*