<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Novelty Search on Eigenform Continuous Learning Tool Tips</title><link>https://www.eigenform.ai/continuous-learning-tool-tips/tags/novelty-search/</link><description>Recent content in Novelty Search on Eigenform Continuous Learning Tool Tips</description><generator>Hugo</generator><language>en-US</language><lastBuildDate>Tue, 17 Nov 2026 09:00:00 +0800</lastBuildDate><atom:link href="https://www.eigenform.ai/continuous-learning-tool-tips/tags/novelty-search/index.xml" rel="self" type="application/rss+xml"/><item><title>Information Gain: Two Gates for Research Agents</title><link>https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/</link><pubDate>Tue, 17 Nov 2026 09:00:00 +0800</pubDate><guid>https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/</guid><description>&lt;figure class="article-image"&gt;
 &lt;picture&gt;
 &lt;source type="image/webp" srcset="https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents_hu_817639235a3841c7.webp 480w, https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents_hu_57d7e5b9adcaaeba.webp 800w, https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents_hu_2c330eadfe75c100.webp 1200w" sizes="(min-width: 900px) 800px, 100vw"&gt;
 &lt;source type="image/webp" srcset="https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents_hu_ffb3ead8df7e05e0.webp 480w, https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents_hu_4c428177acfbdba2.webp 800w, https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents_hu_32717a64e4029db4.webp 1200w" sizes="(min-width: 900px) 800px, 100vw"&gt;
 &lt;img src="https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents_hu_4c428177acfbdba2.webp" alt="Two gates for research agents as an etched parchment engraving: on the left, candidate rule cards fall through a funnel gate with a threshold gauge into a tray of training rows; on the right, near-copies of stored rules bounce off a glass pane into a side bin while one new card passes a second funnel into an archive drawer" title="A rule can train the model without joining the archive: the second gate admits only what it adds" width="1672" height="941" loading="eager" fetchpriority="high" decoding="async" class="zoomable" data-full-url="https://www.eigenform.ai/continuous-learning-tool-tips/information-gain-research-agents/cover-information-gain-research-agents.webp"&gt;
 &lt;/picture&gt;&lt;figcaption&gt;A rule can train the model without joining the archive: the second gate admits only what it adds&lt;/figcaption&gt;&lt;/figure&gt;&lt;p&gt;&lt;strong&gt;TL;DR:&lt;/strong&gt; Reward a hypothesis-generating agent for predictive accuracy alone and it will keep submitting algebraic variants of the first true rule it finds, because each variant is true and each one pays. We solve this by splitting survival into two gates: a standalone signal gate that admits a trajectory into training data when the rule beats a permutation null on held-out pairs net of complexity, and a marginal gate that admits the rule into a persistent archive only when its information gain conditional on the rules already stored has a positive bootstrap lower bound. Rediscoveries train the model and leave the theory unchanged.&lt;/p&gt;</description></item></channel></rss>