<?xml version="1.0" encoding="utf-8"?><feed xmlns="http://www.w3.org/2005/Atom" ><generator uri="https://jekyllrb.com/" version="4.1.1">Jekyll</generator><link href="http://localhost:4000/feed.xml" rel="self" type="application/atom+xml" /><link href="http://localhost:4000/" rel="alternate" type="text/html" /><updated>2025-11-15T23:14:48-08:00</updated><id>http://localhost:4000/feed.xml</id><title type="html">Stavros Diolatzis</title><subtitle>Research Scientist @ Intel</subtitle><entry><title type="html">Ndg</title><link href="http://localhost:4000/2024/09/24/ndg.html" rel="alternate" type="text/html" title="Ndg" /><published>2024-09-24T00:00:00-07:00</published><updated>2024-09-24T00:00:00-07:00</updated><id>http://localhost:4000/2024/09/24/ndg</id><content type="html" xml:base="http://localhost:4000/2024/09/24/ndg.html">&lt;div class=&quot;jekyll-twitter-plugin&quot;&gt;&lt;blockquote class=&quot;twitter-tweet&quot; data-width=&quot;500&quot;&gt;&lt;p lang=&quot;en&quot; dir=&quot;ltr&quot;&gt;Time to share our &lt;a href=&quot;https://twitter.com/hashtag/SIGGRAPH2024?src=hash&amp;amp;ref_src=twsrc%5Etfw&quot;&gt;#SIGGRAPH2024&lt;/a&gt; paper &amp;quot;N-Dimensional Gaussians for Fitting of High Dimensional Functions&amp;quot;!  &lt;br /&gt;&lt;br /&gt;We show that N-Dimensional Gaussian mixtures can be optimized to handle high dimensional inputs like MLPs with a fraction of the training times.&lt;a href=&quot;https://t.co/4edn0ZFwPr&quot;&gt;https://t.co/4edn0ZFwPr&lt;/a&gt; &lt;a href=&quot;https://t.co/d4XKHms2rh&quot;&gt;pic.twitter.com/d4XKHms2rh&lt;/a&gt;&lt;/p&gt;&amp;mdash; Stavros Diolatzis (@StavrosDiol) &lt;a href=&quot;https://twitter.com/StavrosDiol/status/1791438381806702864?ref_src=twsrc%5Etfw&quot;&gt;May 17, 2024&lt;/a&gt;&lt;/blockquote&gt;
&lt;script async=&quot;&quot; src=&quot;https://platform.twitter.com/widgets.js&quot; charset=&quot;utf-8&quot;&gt;&lt;/script&gt;

&lt;/div&gt;</content><author><name></name></author><summary type="html">Time to share our #SIGGRAPH2024 paper &amp;quot;N-Dimensional Gaussians for Fitting of High Dimensional Functions&amp;quot;! We show that N-Dimensional Gaussian mixtures can be optimized to handle high dimensional inputs like MLPs with a fraction of the training times.https://t.co/4edn0ZFwPr pic.twitter.com/d4XKHms2rh&amp;mdash; Stavros Diolatzis (@StavrosDiol) May 17, 2024</summary></entry><entry><title type="html">Gancars</title><link href="http://localhost:4000/2024/02/28/gancars.html" rel="alternate" type="text/html" title="Gancars" /><published>2024-02-28T00:00:00-08:00</published><updated>2024-02-28T00:00:00-08:00</updated><id>http://localhost:4000/2024/02/28/gancars</id><content type="html" xml:base="http://localhost:4000/2024/02/28/gancars.html">&lt;div class=&quot;jekyll-twitter-plugin&quot;&gt;&lt;blockquote class=&quot;twitter-tweet&quot; data-width=&quot;500&quot;&gt;&lt;p lang=&quot;en&quot; dir=&quot;ltr&quot;&gt;Had a great time working with Nicol├ís and everyone else on bringing physically based reflections in 3D GANs in this project! Check it out in Eurographics 2024! &lt;a href=&quot;https://t.co/olGeyjdA34&quot;&gt;https://t.co/olGeyjdA34&lt;/a&gt;&lt;/p&gt;&amp;mdash; Stavros Diolatzis (@StavrosDiol) &lt;a href=&quot;https://twitter.com/StavrosDiol/status/1762843015175962705?ref_src=twsrc%5Etfw&quot;&gt;February 28, 2024&lt;/a&gt;&lt;/blockquote&gt;
&lt;script async=&quot;&quot; src=&quot;https://platform.twitter.com/widgets.js&quot; charset=&quot;utf-8&quot;&gt;&lt;/script&gt;

&lt;/div&gt;</content><author><name></name></author><summary type="html">Had a great time working with Nicol├ís and everyone else on bringing physically based reflections in 3D GANs in this project! Check it out in Eurographics 2024! https://t.co/olGeyjdA34&amp;mdash; Stavros Diolatzis (@StavrosDiol) February 28, 2024</summary></entry><entry><title type="html">Mesogan</title><link href="http://localhost:4000/2023/05/25/mesogan.html" rel="alternate" type="text/html" title="Mesogan" /><published>2023-05-25T00:00:00-07:00</published><updated>2023-05-25T00:00:00-07:00</updated><id>http://localhost:4000/2023/05/25/mesogan</id><content type="html" xml:base="http://localhost:4000/2023/05/25/mesogan.html">&lt;div class=&quot;jekyll-twitter-plugin&quot;&gt;&lt;blockquote class=&quot;twitter-tweet&quot; data-width=&quot;500&quot;&gt;&lt;p lang=&quot;en&quot; dir=&quot;ltr&quot;&gt;I can finally share our CGF paper MesoGAN! We train a 3D GAN with a novel representation to allow generation of arbitrary extent 3D textures. We use these textures with shell mapping  to create the appearance of mesoscale materials like fur or grass. More: &lt;a href=&quot;https://t.co/BCLIXh0Am4&quot;&gt;https://t.co/BCLIXh0Am4&lt;/a&gt; &lt;a href=&quot;https://t.co/T3cSWHk3TS&quot;&gt;pic.twitter.com/T3cSWHk3TS&lt;/a&gt;&lt;/p&gt;&amp;mdash; Stavros Diolatzis (@StavrosDiol) &lt;a href=&quot;https://twitter.com/StavrosDiol/status/1661778040244908044?ref_src=twsrc%5Etfw&quot;&gt;May 25, 2023&lt;/a&gt;&lt;/blockquote&gt;
&lt;script async=&quot;&quot; src=&quot;https://platform.twitter.com/widgets.js&quot; charset=&quot;utf-8&quot;&gt;&lt;/script&gt;

&lt;/div&gt;</content><author><name></name></author><summary type="html">I can finally share our CGF paper MesoGAN! We train a 3D GAN with a novel representation to allow generation of arbitrary extent 3D textures. We use these textures with shell mapping to create the appearance of mesoscale materials like fur or grass. More: https://t.co/BCLIXh0Am4 pic.twitter.com/T3cSWHk3TS&amp;mdash; Stavros Diolatzis (@StavrosDiol) May 25, 2023</summary></entry><entry><title type="html">Nerfshop</title><link href="http://localhost:4000/2023/03/13/nerfshop.html" rel="alternate" type="text/html" title="Nerfshop" /><published>2023-03-13T00:00:00-07:00</published><updated>2023-03-13T00:00:00-07:00</updated><id>http://localhost:4000/2023/03/13/nerfshop</id><content type="html" xml:base="http://localhost:4000/2023/03/13/nerfshop.html">&lt;div class=&quot;jekyll-twitter-plugin&quot;&gt;&lt;blockquote class=&quot;twitter-tweet&quot; data-width=&quot;500&quot;&gt;&lt;p lang=&quot;en&quot; dir=&quot;ltr&quot;&gt;Was a pleasure working with &lt;a href=&quot;https://twitter.com/clementjbn?ref_src=twsrc%5Etfw&quot;&gt;@clementjbn&lt;/a&gt; on this project bringing interactive editing to NeRFs! Keep an eye on this guy! Also thanks to &lt;a href=&quot;https://twitter.com/smallfly?ref_src=twsrc%5Etfw&quot;&gt;@smallfly&lt;/a&gt; for providing one of his great scenes for testing! Check the paper and page for more &lt;a href=&quot;https://t.co/mjmr0S27e1&quot;&gt;https://t.co/mjmr0S27e1&lt;/a&gt;&lt;/p&gt;&amp;mdash; Stavros Diolatzis (@StavrosDiol) &lt;a href=&quot;https://twitter.com/StavrosDiol/status/1635305417059475456?ref_src=twsrc%5Etfw&quot;&gt;March 13, 2023&lt;/a&gt;&lt;/blockquote&gt;
&lt;script async=&quot;&quot; src=&quot;https://platform.twitter.com/widgets.js&quot; charset=&quot;utf-8&quot;&gt;&lt;/script&gt;

&lt;/div&gt;</content><author><name></name></author><summary type="html">Was a pleasure working with @clementjbn on this project bringing interactive editing to NeRFs! Keep an eye on this guy! Also thanks to @smallfly for providing one of his great scenes for testing! Check the paper and page for more https://t.co/mjmr0S27e1&amp;mdash; Stavros Diolatzis (@StavrosDiol) March 13, 2023</summary></entry><entry><title type="html">Siggraph</title><link href="http://localhost:4000/2022/08/12/siggraph.html" rel="alternate" type="text/html" title="Siggraph" /><published>2022-08-12T00:00:00-07:00</published><updated>2022-08-12T00:00:00-07:00</updated><id>http://localhost:4000/2022/08/12/siggraph</id><content type="html" xml:base="http://localhost:4000/2022/08/12/siggraph.html">&lt;div class=&quot;jekyll-twitter-plugin&quot;&gt;&lt;blockquote class=&quot;twitter-tweet&quot; data-width=&quot;500&quot;&gt;&lt;p lang=&quot;en&quot; dir=&quot;ltr&quot;&gt;Leaving &lt;a href=&quot;https://twitter.com/hashtag/SIGGRAPH2022?src=hash&amp;amp;ref_src=twsrc%5Etfw&quot;&gt;#SIGGRAPH2022&lt;/a&gt;, it has been amazing to experience it in person and talk to you all! Hope to see you next year! &lt;a href=&quot;https://twitter.com/emxtyu?ref_src=twsrc%5Etfw&quot;&gt;@emxtyu&lt;/a&gt; &lt;a href=&quot;https://twitter.com/FelixHahnlein?ref_src=twsrc%5Etfw&quot;&gt;@FelixHahnlein&lt;/a&gt; &lt;a href=&quot;https://t.co/fZ4ssTc1JJ&quot;&gt;pic.twitter.com/fZ4ssTc1JJ&lt;/a&gt;&lt;/p&gt;&amp;mdash; Stavros Diolatzis (@StavrosDiol) &lt;a href=&quot;https://twitter.com/StavrosDiol/status/1557967269606543360?ref_src=twsrc%5Etfw&quot;&gt;August 12, 2022&lt;/a&gt;&lt;/blockquote&gt;
&lt;script async=&quot;&quot; src=&quot;https://platform.twitter.com/widgets.js&quot; charset=&quot;utf-8&quot;&gt;&lt;/script&gt;
&lt;/div&gt;</content><author><name></name></author><summary type="html">Leaving #SIGGRAPH2022, it has been amazing to experience it in person and talk to you all! Hope to see you next year! @emxtyu @FelixHahnlein pic.twitter.com/fZ4ssTc1JJ&amp;mdash; Stavros Diolatzis (@StavrosDiol) August 12, 2022</summary></entry><entry><title type="html">Defense</title><link href="http://localhost:4000/2022/06/24/defense.html" rel="alternate" type="text/html" title="Defense" /><published>2022-06-24T00:00:00-07:00</published><updated>2022-06-24T00:00:00-07:00</updated><id>http://localhost:4000/2022/06/24/defense</id><content type="html" xml:base="http://localhost:4000/2022/06/24/defense.html">&lt;div class=&quot;jekyll-twitter-plugin&quot;&gt;&lt;blockquote class=&quot;twitter-tweet&quot; data-width=&quot;500&quot;&gt;&lt;p lang=&quot;en&quot; dir=&quot;ltr&quot;&gt;A few days ago I successfully defended my PhD thesis! Many thanks to my supervisor George Drettakis and the committee: Matthias Zwicker, Hendrik Lensch, Tobias Ritschel and Fr├⌐d├⌐ric Precioso. Now time for some vacations! ΓÿÇ∩╕Å≡ƒîè &lt;a href=&quot;https://t.co/aK6OkCtuFj&quot;&gt;pic.twitter.com/aK6OkCtuFj&lt;/a&gt;&lt;/p&gt;&amp;mdash; Stavros Diolatzis (@StavrosDiol) &lt;a href=&quot;https://twitter.com/StavrosDiol/status/1544642461615771648?ref_src=twsrc%5Etfw&quot;&gt;July 6, 2022&lt;/a&gt;&lt;/blockquote&gt;
&lt;script async=&quot;&quot; src=&quot;https://platform.twitter.com/widgets.js&quot; charset=&quot;utf-8&quot;&gt;&lt;/script&gt;
&lt;/div&gt;</content><author><name></name></author><summary type="html">A few days ago I successfully defended my PhD thesis! Many thanks to my supervisor George Drettakis and the committee: Matthias Zwicker, Hendrik Lensch, Tobias Ritschel and Fr├⌐d├⌐ric Precioso. Now time for some vacations! ΓÿÇ∩╕Å≡ƒîè pic.twitter.com/aK6OkCtuFj&amp;mdash; Stavros Diolatzis (@StavrosDiol) July 6, 2022</summary></entry><entry><title type="html">Active</title><link href="http://localhost:4000/2022/03/02/active.html" rel="alternate" type="text/html" title="Active" /><published>2022-03-02T00:00:00-08:00</published><updated>2022-03-02T00:00:00-08:00</updated><id>http://localhost:4000/2022/03/02/active</id><content type="html" xml:base="http://localhost:4000/2022/03/02/active.html">&lt;div class=&quot;jekyll-twitter-plugin&quot;&gt;&lt;blockquote class=&quot;twitter-tweet&quot; data-width=&quot;500&quot;&gt;&lt;p lang=&quot;en&quot; dir=&quot;ltr&quot;&gt;I am happy to finally share our (&lt;a href=&quot;https://twitter.com/JulienPhilip2?ref_src=twsrc%5Etfw&quot;&gt;@JulienPhilip2&lt;/a&gt; and George Drettakis) ACM TOG paper, to be presented at SIGGRAPH 2022. We propose Active Exploration, an extension of Active Learning using MCMC, for Neural Global Illumination of Variable Scenes. Website: &lt;a href=&quot;https://t.co/FBZC0FWP8m&quot;&gt;https://t.co/FBZC0FWP8m&lt;/a&gt; &lt;a href=&quot;https://t.co/nMciN3AHVz&quot;&gt;pic.twitter.com/nMciN3AHVz&lt;/a&gt;&lt;/p&gt;&amp;mdash; Stavros Diolatzis (@StavrosDiol) &lt;a href=&quot;https://twitter.com/StavrosDiol/status/1499724960729911297?ref_src=twsrc%5Etfw&quot;&gt;March 4, 2022&lt;/a&gt;&lt;/blockquote&gt;
&lt;script async=&quot;&quot; src=&quot;https://platform.twitter.com/widgets.js&quot; charset=&quot;utf-8&quot;&gt;&lt;/script&gt;
&lt;/div&gt;</content><author><name></name></author><summary type="html">I am happy to finally share our (@JulienPhilip2 and George Drettakis) ACM TOG paper, to be presented at SIGGRAPH 2022. We propose Active Exploration, an extension of Active Learning using MCMC, for Neural Global Illumination of Variable Scenes. Website: https://t.co/FBZC0FWP8m pic.twitter.com/nMciN3AHVz&amp;mdash; Stavros Diolatzis (@StavrosDiol) March 4, 2022</summary></entry></feed>