<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Trajectory on 3rd layer</title><link>https://3rdlayer.uk/ja/tags/trajectory/</link><description>Recent content in Trajectory on 3rd layer</description><generator>Hugo -- 0.157.0</generator><language>ja-JP</language><lastBuildDate>Sun, 27 Sep 2026 09:00:00 +0900</lastBuildDate><atom:link href="https://3rdlayer.uk/ja/tags/trajectory/index.xml" rel="self" type="application/rss+xml"/><item><title>Trajectory ML 09: 状態が次に動く方向を学ぶ</title><link>https://3rdlayer.uk/ja/posts/trajectory-ml-09-velocity/</link><pubDate>Sun, 27 Sep 2026 09:00:00 +0900</pubDate><guid>https://3rdlayer.uk/ja/posts/trajectory-ml-09-velocity/</guid><description>観測ペアからベクトル場を推定し、1ステップ予測と経路全体の予測の違いを検証する。</description></item><item><title>Trajectory ML 10: 最適輸送で集団の移動を考える</title><link>https://3rdlayer.uk/ja/posts/trajectory-ml-10-transport/</link><pubDate>Sun, 27 Sep 2026 09:00:00 +0900</pubDate><guid>https://3rdlayer.uk/ja/posts/trajectory-ml-10-transport/</guid><description>周辺分布の制約を満たす輸送計画を求め、個体追跡との違いとSinkhorn正則化の役割を理解する。</description></item><item><title>Trajectory ML 11: 連続ダイナミクスとFlow Matching</title><link>https://3rdlayer.uk/ja/posts/trajectory-ml-11-continuous/</link><pubDate>Sun, 27 Sep 2026 09:00:00 +0900</pubDate><guid>https://3rdlayer.uk/ja/posts/trajectory-ml-11-continuous/</guid><description>Neural ODE、SDE、Flow Matchingをつなぎ、生成経路と実際の時間発展を区別する。</description></item><item><title>Trajectory ML 12: 帰結の予測、生存分析、因果推論</title><link>https://3rdlayer.uk/ja/posts/trajectory-ml-12-fate-causality/</link><pubDate>Sun, 27 Sep 2026 09:00:00 +0900</pubDate><guid>https://3rdlayer.uk/ja/posts/trajectory-ml-12-fate-causality/</guid><description>打ち切りを扱い、初期状態から結末を予測し、予測上の関連と介入の効果を区別する。</description></item></channel></rss>