<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>UI理解 on AI 早报</title><link>https://ai-news.example.com/tags/ui%E7%90%86%E8%A7%A3/</link><description>Recent content in UI理解 on AI 早报</description><generator>Hugo</generator><language>zh-CN</language><lastBuildDate>Sat, 04 Jul 2026 06:39:25 +0800</lastBuildDate><atom:link href="https://ai-news.example.com/tags/ui%E7%90%86%E8%A7%A3/feed.xml" rel="self" type="application/rss+xml"/><item><title>[EN] Google推出ScreenAI：端到端视觉语言模型，统一理解UI界面与信息图表</title><link>https://ai-news.example.com/articles/en-googlescreenaiui/</link><pubDate>Sat, 04 Jul 2026 06:39:25 +0800</pubDate><guid>https://ai-news.example.com/articles/en-googlescreenaiui/</guid><description>Google Research发布ScreenAI，一种基于PaLI架构并融合pix2struct灵活分块策略的5B参数视觉语言模型，专为理解用户界面与信息图表设计。通过自监督预训练和屏幕标注任务，ScreenAI在WebSRC、MoTIF、ChartQA等多项基准测试中达到或超越同规模模型的最优性能，同时发布了三个新数据集。</description></item></channel></rss>