Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
GUI-Rise: Structured Reasoning and History Summarization for GUI Navigation
Authors: Tao Liu, Chongyu Wang, Rongjie Li, Yingchen Yu, Xuming He, Song Bai
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Comprehensive evaluations on standard benchmarks demonstrate state-of-the-art results under identical training data conditions, with particularly strong performance in out-of-domain scenarios. |
| Researcher Affiliation | Collaboration | 1Shanghai Tech University 2Byte Dance 3Shanghai Engineering Research Center of Intelligent Vision and Imaging EMAIL EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes methods through architectural diagrams (Figure 1, Figure 2) and textual descriptions in Section 3 and 4, but does not include explicit pseudocode or algorithm blocks. |
| Open Source Code | Yes | Code is available at https://leon022.github.io/GUI-Rise. |
| Open Datasets | Yes | Datasets. Our experiments use three offline GUI navigation benchmarks and one online benchmark. (i) Mind2Web (Web) [9], (ii) AITW (Mobile) [30], (iii) GUIAct [6], (iv) Mini Wob (Online) [35]. |
| Dataset Splits | Yes | Mind2Web is evaluated using element accuracy (Ele.Acc), operation F1 (Op.F1), and step success rate (Step SR) across three verified test splits test-task, test-website, and test-domain covering variations in tasks, websites, and domains. For in-domain evaluation, we train the model on the training set of AITW and test it on the respective test set. |
| Hardware Specification | No | The paper mentions running experiments on the HPC Platform of Shanghai Tech University in the acknowledgments, but specific hardware details such as GPU/CPU models or memory specifications are not provided in the experimental setup or any other section. |
| Software Dependencies | No | The paper mentions using specific models and algorithms like the Qwen-VL series [38, 3], GPT-4o-mini [52], and the GRPO [33] algorithm. However, it does not explicitly provide details on software dependencies such as programming languages, libraries, or frameworks with their specific version numbers. |
| Experiment Setup | No | The paper describes the two-stage training strategy, the use of GRPO with three reward functions, and the high-level methodology. However, specific experimental setup details such as concrete hyperparameter values (e.g., learning rate, batch size, number of epochs), optimizer settings, or model initialization details are not explicitly provided in the main body of the paper. |