---
格式版本: 2
标题: "SkillLens: Visual Skill Cards for Retrieval-Augmented GUI Action Prediction and On-Policy Distillation"
原文链接: "https://arxiv.org/abs/2608.10775"
发布日期: "2026-08-11"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Tue, 11 Aug 2026 10:28:59 UTC (1,975 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-12T11:04:31+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-12T11:03:39+08:00"
入库时间: "2026-08-12T03:04:31.166Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=SUE&searchtype=all"
匹配关键词:
  - "SUE"
  - "AI"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 5
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "论文主题为GUI动作预测与强化学习，与超节点/AI Rack/机柜级AI基础设施完全无关，无任何技术或商业信号。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-12T11:04:38+08:00"
AI主题相关性: 0
AI来源权威性: 5
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
采集批次: "2026年8月12日2点04分20秒"
采集批次ID: "20260812-020420-289"
去重键: "https://arxiv.org/abs/2608.10775"
---

## Computer Science > Artificial Intelligence

## Title:SkillLens: Visual Skill Cards for Retrieval-Augmented GUI Action Prediction and On-Policy Distillation

Authors:[Zhou Liu](https://arxiv.org/search/cs?searchtype=author&query=Liu,+Z), [Ligang Huang](https://arxiv.org/search/cs?searchtype=author&query=Huang,+L), [Zeli Su](https://arxiv.org/search/cs?searchtype=author&query=Su,+Z), [Zewei Pan](https://arxiv.org/search/cs?searchtype=author&query=Pan,+Z), [Zhaoyang Han](https://arxiv.org/search/cs?searchtype=author&query=Han,+Z), [Xing Chen](https://arxiv.org/search/cs?searchtype=author&query=Chen,+X), [Yuanfeng Song](https://arxiv.org/search/cs?searchtype=author&query=Song,+Y), [Wentao Zhang](https://arxiv.org/search/cs?searchtype=author&query=Zhang,+W)

[View PDF](https://arxiv.org/pdf/2608.10775) [HTML (experimental)](https://arxiv.org/html/2608.10775v1)

> Abstract:Computer-using agents can perceive rich software interfaces, yet their decisions often lack visual procedural memory: they may recognize individual controls without identifying which familiar workflow is active, which control matters next, or what evidence would confirm progress. Raw interaction traces preserve such information but are long and noisy to condition on, whereas text-only skills often omit the visual state that makes a procedure applicable. We introduce Visual Skill Cards (VSCs), a state-conditioned memory representation that binds reusable procedures with applicability cues, visual evidence, and verification signals. SkillLens constructs VSCs from heterogeneous interaction experience through Trace-to-Visual-Skill-Card and, at inference time, retrieves relevant cards and selectively expands only the evidence needed by a fixed visual-language model executor for grounded GUI action prediction. The same representation also supports CardDistill, which uses VSC evidence as privileged teacher context to train a student that acts without runtime card retrieval. Across Multimodal-Mind2Web and WebLINX-BrowserGym, SkillLens improves the frozen GPT-5.4-mini executor by +11.6 points in Step SR and +2.9 points in Overall, respectively; CardDistill further improves the corresponding student-only Qwen3-VL-2B metrics by +12.0 and +3.2 points.

| Subjects: | Artificial Intelligence (cs.AI) |
| --- | --- |
| Cite as: | [arXiv:2608.10775](https://arxiv.org/abs/2608.10775) \[cs.AI\] |
|  | (or [arXiv:2608.10775v1](https://arxiv.org/abs/2608.10775v1) \[cs.AI\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.10775](https://doi.org/10.48550/arXiv.2608.10775) |

## Submission history

From: Liu Zhou \[[view email](https://arxiv.org/show-email/2efc77d5/2608.10775)\]  
**\[v1\]** Tue, 11 Aug 2026 10:28:59 UTC (1,975 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.10775) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))