<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>文章 on syhdtc668 · Research &amp; Engineering Notes</title><link>https://syhdtc668.top/posts/</link><description>Recent content in 文章 on syhdtc668 · Research &amp; Engineering Notes</description><generator>Hugo</generator><language>zh-CN</language><copyright>© 2024–2026 syhdtc668</copyright><lastBuildDate>Mon, 31 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://syhdtc668.top/posts/index.xml" rel="self" type="application/rss+xml"/><item><title>两篇论文中的球面锥点云表示：Flying on Point Clouds 与 AERO-MPPI</title><link>https://syhdtc668.top/writing/multi-resolution-spherical-cone-partition/</link><pubDate>Mon, 31 Aug 2026 00:00:00 +0000</pubDate><guid>https://syhdtc668.top/writing/multi-resolution-spherical-cone-partition/</guid><description>逐项区分论文原文、公式推导与工程推断，比较两种面向无人机 LiDAR 的球面锥表示。</description></item><item><title>Windows 睡眠、休眠与电源计划：一份可验证的排查指南</title><link>https://syhdtc668.top/writing/windows-sleep-and-powercfg/</link><pubDate>Sun, 30 Aug 2026 00:00:00 +0000</pubDate><guid>https://syhdtc668.top/writing/windows-sleep-and-powercfg/</guid><description>不依赖注册表切换电源模型：先确认硬件支持的睡眠状态，再针对合盖、唤醒来源、诊断报告和电源计划逐项验证。</description></item><item><title>从 QKᵀ 到加权聚合：用 3 个 Agent 直观理解 Attention</title><link>https://syhdtc668.top/writing/attention-qkv-bmm/</link><pubDate>Sun, 30 Aug 2026 00:00:00 +0000</pubDate><guid>https://syhdtc668.top/writing/attention-qkv-bmm/</guid><description>用三个 Agent 的具体数值，逐步拆解 QKᵀ 相似度和注意力加权聚合。</description></item></channel></rss>