Qinzhuo Wu
Research Subjects
Qinzhuo Wu has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Jian Luan
MiLM Plus Xiaomi Inc Beijing China
9 shared publications - Wei Liu
MiLM Plus Xiaomi Inc
8 shared publications - Bin Wang
MiLM Plus Xiaomi Inc Beijing China
4 shared publications - Pengzhi Gao
Xiaomi Inc Beijing China
4 shared publications Ang Li
1 shared publication- Hanhao Li1 shared publication
Latest Publications
- Open Access
- 5 November 2026
- Published by MDPI AG in Preprints.org
World Models for Digital Agents: An Agent-Centric Survey
Digital agents operating in software-defined environments, including games, web interfaces, mobile GUIs, tool-use systems, and code repositories, often benefit from anticipating the consequences of actions before execution. World models provide a principled mechanism for such anticipation, but their utility in digital settings is neither automatic nor uniform. Unlike physical environments, digital transitions are interface-mediated and program-governed, and the value of prediction depends less on perceptual realism than on preserving functionally relevant consequences for downstream decisions. This survey presents the first dedicated agent-centric treatment of world models across major digital-agent domains. We introduce a unified design space, W = (X , L, U), organized around three complementary questions: what is modeled, how predictive capability is built, and how world models are used inside the agent loop. Through this lens, we synthesize a rapidly growing literature spanning semantic transition models, visual and latent simulators, tool-state predictors, executable code-based world models, and hybrid predictive systems. We further examine environments, datasets, and evaluation protocols as a coupled empirical pipeline, and highlight open challenges, notably the persistent gap between predictive fidelity and downstream utility.
- 5 November 2026
- Published by MDPI AG in Preprints.org
Open AccessWorld Models for Digital Agents: An Agent-Centric Survey
Digital agents operating in software-defined environments, including games, web interfaces, mobile GUIs, tool-use systems, and code repositories, often benefit from anticipating the consequences of actions before execution. World models provide a principled mechanism for such anticipation, but their utility in digital settings is neither automatic nor uniform. Unlike physical environments, digital transitions are interface-mediated and program-governed, and the value of prediction depends less on perceptual realism than on preserving functionally relevant consequences for downstream decisions. This survey presents the first dedicated agent-centric treatment of world models across major digital-agent domains. We introduce a unified design space, W = (X , L, U), organized around three complementary questions: what is modeled, how predictive capability is built, and how world models are used inside the agent loop. Through this lens, we synthesize a rapidly growing literature spanning semantic transition models, visual and latent simulators, tool-state predictors, executable code-based world models, and hybrid predictive systems. We further examine environments, datasets, and evaluation protocols as a coupled empirical pipeline, and highlight open challenges, notably the persistent gap between predictive fidelity and downstream utility.
- Open Access
MobileBench-OL: A Comprehensive Chinese Benchmark for Evaluating Mobile GUI Agents in Real-World Environment
Recent advances in mobile Graphical User Interface (GUI) agents highlight the growing need for comprehensive evaluation benchmarks. While new online benchmarks offer more realistic testing than offline ones, they tend to focus on the agents' task instruction-following ability while neglecting their reasoning and exploration ability. Moreover, these benchmarks do not consider the random noise in real-world mobile environments. This leads to a gap between benchmarks and real-world environments. To addressing these limitations, we propose MobileBench-OL, an online benchmark with 1080 tasks from 80 Chinese apps. It measures task execution, complex reasoning, and noise robustness of agents by including 5 subsets, which set multiple evaluation dimensions. We also provide an auto-eval framework with a reset mechanism, enabling stable and repeatable real-world benchmarking. Evaluating 12 leading GUI agents on MobileBench-OL shows significant room for improvement to meet real-world requirements. Human evaluation further confirms that MobileBench-OL can reliably measure the performance of leading GUI agents in real environments. Our data and code will be released upon acceptance.
Open AccessMobileBench-OL: A Comprehensive Chinese Benchmark for Evaluating Mobile GUI Agents in Real-World Environment
Recent advances in mobile Graphical User Interface (GUI) agents highlight the growing need for comprehensive evaluation benchmarks. While new online benchmarks offer more realistic testing than offline ones, they tend to focus on the agents' task instruction-following ability while neglecting their reasoning and exploration ability. Moreover, these benchmarks do not consider the random noise in real-world mobile environments. This leads to a gap between benchmarks and real-world environments. To addressing these limitations, we propose MobileBench-OL, an online benchmark with 1080 tasks from 80 Chinese apps. It measures task execution, complex reasoning, and noise robustness of agents by including 5 subsets, which set multiple evaluation dimensions. We also provide an auto-eval framework with a reset mechanism, enabling stable and repeatable real-world benchmarking. Evaluating 12 leading GUI agents on MobileBench-OL shows significant room for improvement to meet real-world requirements. Human evaluation further confirms that MobileBench-OL can reliably measure the performance of leading GUI agents in real environments. Our data and code will be released upon acceptance.
- Open Access
BacktrackAgent: Enhancing GUI Agent with Error Detection and Backtracking Mechanism
Graphical User Interface (GUI) agents have gained substantial attention due to their impressive capabilities to complete tasks through multiple interactions within GUI environments. However, existing agents primarily focus on enhancing the accuracy of individual actions and often lack effective mechanisms for detecting and recovering from errors. To address these shortcomings, we propose the BacktrackAgent, a robust framework that incorporates a backtracking mechanism to improve task completion efficiency. BacktrackAgent includes verifier, judger, and reflector components as modules for error detection and recovery, while also applying judgment rewards to further enhance the agent's performance. Additionally, we develop a training dataset specifically designed for the backtracking mechanism, which considers the outcome pages after action executions. Experimental results show that BacktrackAgent has achieved performance improvements in both task success rate and step accuracy on Mobile3M and Auto-UI benchmarks. Our data and code will be released upon acceptance.
Open AccessBacktrackAgent: Enhancing GUI Agent with Error Detection and Backtracking Mechanism
Graphical User Interface (GUI) agents have gained substantial attention due to their impressive capabilities to complete tasks through multiple interactions within GUI environments. However, existing agents primarily focus on enhancing the accuracy of individual actions and often lack effective mechanisms for detecting and recovering from errors. To address these shortcomings, we propose the BacktrackAgent, a robust framework that incorporates a backtracking mechanism to improve task completion efficiency. BacktrackAgent includes verifier, judger, and reflector components as modules for error detection and recovery, while also applying judgment rewards to further enhance the agent's performance. Additionally, we develop a training dataset specifically designed for the backtracking mechanism, which considers the outcome pages after action executions. Experimental results show that BacktrackAgent has achieved performance improvements in both task success rate and step accuracy on Mobile3M and Auto-UI benchmarks. Our data and code will be released upon acceptance.
- Open Access
ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation
Recently, mobile AI agents have gained increasing attention. Given a task, mobile AI agents can interact with mobile devices in multiple steps and finally form a GUI flow that solves the task. However, existing agents tend to focus on most task-relevant elements at each step, leading to local optimal solutions and ignoring the overall GUI flow. To address this issue, we constructed a training dataset called MobileReach, which breaks the task into page reaching and operation subtasks. Furthermore, we propose ReachAgent, a two-stage framework that focuses on improving its task-completion abilities. It utilizes the page reaching and page operation subtasks, along with reward-based preference GUI flows, to further enhance the agent. Experimental results show that ReachAgent significantly improves the IoU Acc and Text Acc by 7.12% and 7.69% on the step-level and 4.72% and 4.63% on the task-level compared to the SOTA agent. Our data and code will be released upon acceptance.
Open AccessReachAgent: Enhancing Mobile Agent via Page Reaching and Operation
Recently, mobile AI agents have gained increasing attention. Given a task, mobile AI agents can interact with mobile devices in multiple steps and finally form a GUI flow that solves the task. However, existing agents tend to focus on most task-relevant elements at each step, leading to local optimal solutions and ignoring the overall GUI flow. To address this issue, we constructed a training dataset called MobileReach, which breaks the task into page reaching and operation subtasks. Furthermore, we propose ReachAgent, a two-stage framework that focuses on improving its task-completion abilities. It utilizes the page reaching and page operation subtasks, along with reward-based preference GUI flows, to further enhance the agent. Experimental results show that ReachAgent significantly improves the IoU Acc and Text Acc by 7.12% and 7.69% on the step-level and 4.72% and 4.63% on the task-level compared to the SOTA agent. Our data and code will be released upon acceptance.
- Open Access
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 5601-5614
SPO: Self Preference Optimization with Self Regularization
Yuhao Sun, Yifan Zhang, Quandong Wang, Qinzhuo Wu, Wei Liu, Jian Luan. Findings of the Association for Computational Linguistics: EMNLP 2025. 2025.
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 5601-5614
Open AccessSPO: Self Preference Optimization with Self Regularization
Yuhao Sun, Yifan Zhang, Quandong Wang, Qinzhuo Wu, Wei Liu, Jian Luan. Findings of the Association for Computational Linguistics: EMNLP 2025. 2025.
- Open Access
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 4250-4272
BacktrackAgent: Enhancing GUI Agent with Error Detection and Backtracking Mechanism
Qinzhuo Wu, Pengzhi Gao, Wei Liu, Jian Luan. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 4250-4272
Open AccessBacktrackAgent: Enhancing GUI Agent with Error Detection and Backtracking Mechanism
Qinzhuo Wu, Pengzhi Gao, Wei Liu, Jian Luan. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.
- Open Access
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 4760-4775
ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation
Qinzhuo Wu, Wei Liu, Jian Luan, Bin Wang. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 4760-4775
Open AccessReachAgent: Enhancing Mobile Agent via Page Reaching and Operation
Qinzhuo Wu, Wei Liu, Jian Luan, Bin Wang. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.
- Top Cited PapersOpen Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 10231-10251
MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding
Qinzhuo Wu, Weikai Xu, Wei Liu, Tao Tan, Liujian Liujianfeng, Ang Li, Jian Luan, Bin Wang, Shuo Shang. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 10231-10251
Top Cited PapersOpen AccessMobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding
Qinzhuo Wu, Weikai Xu, Wei Liu, Tao Tan, Liujian Liujianfeng, Ang Li, Jian Luan, Bin Wang, Shuo Shang. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 18315-18339
ToolPlanner: A Tool Augmented LLM for Multi Granularity Instructions with Path Planning and Feedback
Qinzhuo Wu, Wei Liu, Jian Luan, Bin Wang. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 18315-18339
Open AccessToolPlanner: A Tool Augmented LLM for Multi Granularity Instructions with Path Planning and Feedback
Qinzhuo Wu, Wei Liu, Jian Luan, Bin Wang. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.