<?xml version="1.0" encoding="utf-8" standalone="yes" ?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Foundation Models | Timon Weitkamp | No Worries, Just Maps</title>
    <link>https://timonweitkamp.me/tags/foundation-models/</link>
      <atom:link href="https://timonweitkamp.me/tags/foundation-models/index.xml" rel="self" type="application/rss+xml" />
    <description>Foundation Models</description>
    <generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Wed, 28 Oct 2026 00:00:00 +0000</lastBuildDate>
    <image>
      <url>https://timonweitkamp.me/media/icon_hu11869890484969746898.png</url>
      <title>Foundation Models</title>
      <link>https://timonweitkamp.me/tags/foundation-models/</link>
    </image>
    
    <item>
      <title>Satellite embeddings for smallholder irrigation mapping: how far do the models transfer?</title>
      <link>https://timonweitkamp.me/event/waternet-2026/</link>
      <pubDate>Wed, 28 Oct 2026 00:00:00 +0000</pubDate>
      <guid>https://timonweitkamp.me/event/waternet-2026/</guid>
      <description>&lt;p&gt;In my PhD, models for mapping smallholder irrigation transferred badly from one region in Mozambique to another. Geospatial foundation models promise features that work everywhere, so I tested whether that holds for irrigation.&lt;/p&gt;
&lt;p&gt;I compared plain Sentinel-2 with embeddings from two foundation models, AlphaEarth and TESSERA. I trained classifiers in one study area, applied them in another, and also transferred them between years (2021 and 2022), using the field data from my PhD study areas.&lt;/p&gt;
&lt;p&gt;&lt;a href=&#34;https://timonweitkamp.me/project/mapping-smallholder-irrigation/&#34;&gt;More about mapping smallholder irrigation&lt;/a&gt;&lt;/p&gt;
</description>
    </item>
    
  </channel>
</rss>
