Saved in:
| Main Authors: | , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.21799 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908750451834880 |
|---|---|
| author | Mi, Hongze Feng, Yibo Lu, Wenjie Wang, Yuqi Li, Jinyuan Cao, Song Cui, He Tian, Tengfei Zhang, Xuelin Luo, Haotian Sun, Di Fang, Jun Chai, Hua Tan, Naiqiang Pan, Gang |
| author_facet | Mi, Hongze Feng, Yibo Lu, Wenjie Wang, Yuqi Li, Jinyuan Cao, Song Cui, He Tian, Tengfei Zhang, Xuelin Luo, Haotian Sun, Di Fang, Jun Chai, Hua Tan, Naiqiang Pan, Gang |
| contents | Graphical User Interface (GUI) agents aim to automate a wide spectrum of human tasks by emulating user interaction. Despite rapid advancements, current approaches are hindered by several critical challenges: data bottleneck in end-to-end training, high cost of delayed error detection, and risk of contradictory guidance. Inspired by the human cognitive loop of Thinking, Alignment, and Reflection, we present D-Artemis -- a novel deliberative framework in this paper. D-Artemis leverages a fine-grained, app-specific tip retrieval mechanism to inform its decision-making process. It also employs a proactive Pre-execution Alignment stage, where Thought-Action Consistency (TAC) Check module and Action Correction Agent (ACA) work in concert to mitigate the risk of execution failures. A post-execution Status Reflection Agent (SRA) completes the cognitive loop, enabling strategic learning from experience. Crucially, D-Artemis enhances the capabilities of general-purpose Multimodal large language models (MLLMs) for GUI tasks without the need for training on complex trajectory datasets, demonstrating strong generalization. D-Artemis establishes new state-of-the-art (SOTA) results across both major benchmarks, achieving a 75.8% success rate on AndroidWorld and 96.8% on ScreenSpot-V2. Extensive ablation studies further demonstrate the significant contribution of each component to the framework. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_21799 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | D-Artemis: A Deliberative Cognitive Framework for Mobile GUI Multi-Agents Mi, Hongze Feng, Yibo Lu, Wenjie Wang, Yuqi Li, Jinyuan Cao, Song Cui, He Tian, Tengfei Zhang, Xuelin Luo, Haotian Sun, Di Fang, Jun Chai, Hua Tan, Naiqiang Pan, Gang Artificial Intelligence Graphical User Interface (GUI) agents aim to automate a wide spectrum of human tasks by emulating user interaction. Despite rapid advancements, current approaches are hindered by several critical challenges: data bottleneck in end-to-end training, high cost of delayed error detection, and risk of contradictory guidance. Inspired by the human cognitive loop of Thinking, Alignment, and Reflection, we present D-Artemis -- a novel deliberative framework in this paper. D-Artemis leverages a fine-grained, app-specific tip retrieval mechanism to inform its decision-making process. It also employs a proactive Pre-execution Alignment stage, where Thought-Action Consistency (TAC) Check module and Action Correction Agent (ACA) work in concert to mitigate the risk of execution failures. A post-execution Status Reflection Agent (SRA) completes the cognitive loop, enabling strategic learning from experience. Crucially, D-Artemis enhances the capabilities of general-purpose Multimodal large language models (MLLMs) for GUI tasks without the need for training on complex trajectory datasets, demonstrating strong generalization. D-Artemis establishes new state-of-the-art (SOTA) results across both major benchmarks, achieving a 75.8% success rate on AndroidWorld and 96.8% on ScreenSpot-V2. Extensive ablation studies further demonstrate the significant contribution of each component to the framework. |
| title | D-Artemis: A Deliberative Cognitive Framework for Mobile GUI Multi-Agents |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2509.21799 |