<?xml version="1.0"?>
<rss version="2.0">
  <channel>
    <title>robots.net blog for togelius</title>
    <link>http://robots.net/person/togelius/</link>
    <description>robots.net blog for togelius</description>
    <language>en-us</language>
    <generator>mod_virgule</generator>
    <pubDate>Tue, 2 Dec 2008 19:18:27 GMT</pubDate>
    <item>
      <pubDate>Thu, 22 Mar 2007 16:53:13 GMT</pubDate>
      <title>22 Mar 2007</title>
      <link>http://robots.net/person/togelius/diary.html?start=0</link>
      <guid>http://robots.net/person/togelius/diary.html?start=0</guid>
      <description>At Essex, we have for some time been working on &lt;a&#xD;
href="http://togelius.blogspot.com/2006/04/evolutionary-car-racing-videos.html"&gt;automatically&#xD;
learning how to race cars in simulation&lt;/a&gt;. It turns out&#xD;
that a combination of evolutionary algorithms and neural&#xD;
networks can learn how to beat all humans in racing games,&#xD;
and also come up with some quite interesting, novel&#xD;
behaviours, which might one day make their way into&#xD;
commercial racing games. While this is simulation, the race&#xD;
is now on for the real thing - we are setting up &lt;a&#xD;
href="http://www.essex.ac.uk/news/2007/nr20070320.htm"&gt;a&#xD;
competition for AI developers, where the goal is to win a&#xD;
race between model cars on real tracks&lt;/a&gt;. As the cars will&#xD;
be around half a meter long, the cost of participating will&#xD;
be a fraction of that for the famous DARPA Grand Challenge,&#xD;
whereas the challenges will be similar in terms of computer&#xD;
vision and AI.</description>
    </item>
  </channel>
</rss>
