<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>Universal algebra gives universal approximation for neural nets</title>
        <link>https://peertube.wtf/videos/watch/3213f9a9-9721-4927-bb0f-34c37a8c2747</link>
        <description>This is my talk at the University of Rochester's combinatorics seminar for the spring of 2021. I discuss a theorem of Murskiĭ from the 1970s which states that (under some extremely mild assumptions) a randomly-chosen finite algebra is primal with probability 1. I sketch the proof of this result from universal algebra, which is quite combinatorial in nature. I also explain how this relates to the theory of neural nets and gives a discrete universal approximation theorem. Connetions to alien computer systems, a sharp version of Stirling's approximation, operads, and rock-paper-scissors are noted. Slides: https://aten.cool/documents/slides/rochester_combinatorics_seminar.pdf Logic on other planets: https://www.math.hawaii.edu/~jb/planets.pdf Alex Iosevich: https://alexiosevich.com/</description>
        <lastBuildDate>Tue, 18 Aug 2026 12:30:38 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>PeerTube - https://peertube.wtf</generator>
        <image>
            <title>Universal algebra gives universal approximation for neural nets</title>
            <url>https://peertube.wtf/lazy-static/avatars/45856119-df14-4b5f-bff8-4807ae2cfbd2.png</url>
            <link>https://peertube.wtf/videos/watch/3213f9a9-9721-4927-bb0f-34c37a8c2747</link>
        </image>
        <copyright>All rights reserved, unless otherwise specified in the terms specified at https://peertube.wtf/about and potential licenses granted by each content's rightholder.</copyright>
        <atom:link href="https://peertube.wtf/feeds/video-comments.xml?videoId=3213f9a9-9721-4927-bb0f-34c37a8c2747" rel="self" type="application/rss+xml"/>
    </channel>
</rss>