<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Newton &amp; Networks</title><description>Writing on physics-based modelling and AI in heavy industry, by George Drakoulas, PhD.</description><link>https://newtonandnetworks.com/</link><item><title>Why surrogate models beat full simulation in the design loop</title><link>https://newtonandnetworks.com/blog/surrogate-models/</link><guid isPermaLink="true">https://newtonandnetworks.com/blog/surrogate-models/</guid><description>A four-hour simulation run is fine for validation and useless for exploration. What changes when you need an answer in seconds, and where it stops being trustworthy.</description><pubDate>Tue, 15 Sep 2026 00:00:00 GMT</pubDate><category>maritime</category><category>surrogate models</category><category>geometric deep learning</category><category>digital twins</category></item></channel></rss>