Unverified50% confidenceFactExact time
GUI grounding is a foundational task in the GUI agent field, referring to an agent locating the corresponding interactive element in an interface screenshot given a natural language instruction and outputting coordinates to execute a click.
1
Sources
50%
Confidence
Long-term
Relevance
10/26/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe underlying technology of GUI agents relies on the collaborative operation of visual language models' screen comprehension capabilities and action planning modules: first performing semantic understanding of screen screenshots to identify interface elements, then having the planning module generate specific instruction sequences for execution.69% similarUnverifiedGUI Agent通过截图理解屏幕状态并输出鼠标键盘操作指令，无需目标应用提供结构化接口67% similarUnverifiedThe fusion of conversational interaction and traditional GUI may be the most realistic near-term form of Intelligent UI deployment.65% similarUnverifiedIntelligent UI can dynamically generate buttons, sliders, interactive charts, and usable widgets within responses based on conversational intent.63% similarUnverifiedGUI智能体的界面元素识别通常结合目标检测（Object Detection）定位UI组件位置和OCR文字识别读取界面文字两种技术63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1098566API
curl https://kongchang.com/api/v1/knowledge/claims/1098566MCP
get_claim(id=1098566)