<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Tong Xiao · News</title>
  <id>https://xiaotong.me/</id>
  <link href="https://xiaotong.me/"/>
  <link rel="self" href="https://xiaotong.me/feed.xml"/>
  <author><name>Tong Xiao</name></author>
  <updated>2026-10-01T00:00:00Z</updated>
  <entry>
    <id>tag:xiaotong.me,2026:news/2026-10/12</id>
    <title>Robot Overflow is live: a hub where we can better understand the robot industry.</title>
    <link href="https://xiaotong.me/RobotOverflow/"/>
    <updated>2026-10-01T00:00:00Z</updated>
    <content type="html">&lt;a href=&quot;https://xiaotong.me/RobotOverflow/&quot;&gt;Robot Overflow&lt;/a&gt; is live: a hub where we can better understand the robot industry.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2026:news/2026-09/11</id>
    <title>I made Robot Real2Sim to better understand what the Real2Sim system is and its potential value.</title>
    <link href="https://xiaotong.me/Real2Sim/"/>
    <updated>2026-09-01T00:00:00Z</updated>
    <content type="html">I made &lt;a href=&quot;https://xiaotong.me/Real2Sim/&quot;&gt;Robot Real2Sim&lt;/a&gt; to better understand what the Real2Sim system is and its potential value.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2026:news/2026-03/10</id>
    <title>I left Yutori to spend more time with family.</title>
    <link href="https://xiaotong.me/news/"/>
    <updated>2026-03-01T00:00:00Z</updated>
    <content type="html">I left Yutori to spend more time with family.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2025:news/2025-11/9</id>
    <title>We introduced Yutori Navigator, a state-of-the-art AI web agent. I built the RL model for it.</title>
    <link href="https://yutori.com/blog/introducing-navigator"/>
    <updated>2025-11-01T00:00:00Z</updated>
    <content type="html">We introduced &lt;a href=&quot;https://yutori.com/blog/introducing-navigator&quot;&gt;Yutori Navigator&lt;/a&gt;, a state-of-the-art AI web agent. I built the RL model for it.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2025:news/2025-02/8</id>
    <title>I joined Yutori to build AI agents for the web.</title>
    <link href="https://yutori.com"/>
    <updated>2025-02-01T00:00:00Z</updated>
    <content type="html">I joined &lt;a href=&quot;https://yutori.com&quot;&gt;Yutori&lt;/a&gt; to build AI agents for the web.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2024:news/2024-08/7</id>
    <title>We published the Llama 3 paper, and I was excited to contribute to the vision post-training.</title>
    <link href="https://ai.meta.com/research/publications/the-llama-3-herd-of-models/"/>
    <updated>2024-08-01T00:00:00Z</updated>
    <content type="html">We published the &lt;a href=&quot;https://ai.meta.com/research/publications/the-llama-3-herd-of-models/&quot;&gt;Llama 3 paper&lt;/a&gt;, and I was excited to contribute to the vision post-training.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2023:news/2023-12/6</id>
    <title>We shipped Emu Edit, a diffusion model for free-form image edition from text instructions. Happy to lead the production of it.</title>
    <link href="https://emu-edit.metademolab.com/"/>
    <updated>2023-12-01T00:00:00Z</updated>
    <content type="html">We shipped &lt;a href=&quot;https://emu-edit.metademolab.com/&quot;&gt;Emu Edit&lt;/a&gt;, a diffusion model for free-form image edition from text instructions. Happy to lead the production of it.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2022:news/2022-10/5</id>
    <title>We shipped Quest Pro, the world’s first standalone VR headset with face tracking capabilities; see the Movement SDK, customer reviews, and a video of a user having fun with face tracking.</title>
    <link href="https://developer.oculus.com/blog/movement-sdk-sample-github-eye-tracking/"/>
    <updated>2022-10-01T00:00:00Z</updated>
    <content type="html">We shipped &lt;strong&gt;Quest Pro&lt;/strong&gt;, the world’s first standalone VR headset with face tracking capabilities; see the &lt;a href=&quot;https://developer.oculus.com/blog/movement-sdk-sample-github-eye-tracking/&quot;&gt;Movement SDK&lt;/a&gt;, &lt;a href=&quot;https://www.reddit.com/r/VRchat/comments/11z41ey/vrchat_eye_face_tracking_on_quest_pro/&quot;&gt;customer reviews&lt;/a&gt;, and a &lt;a href=&quot;https://www.youtube.com/watch?v=KCutNAwIWfo&quot;&gt;video of a user having fun with face tracking&lt;/a&gt;.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2017:news/2017-03/4</id>
    <title>We released Open-ReID — a deep learning library for person re-identification.</title>
    <link href="https://github.com/Cysu/open-reid"/>
    <updated>2017-03-01T00:00:00Z</updated>
    <content type="html">We released &lt;a href=&quot;https://github.com/Cysu/open-reid&quot;&gt;Open-ReID — a deep learning library for person re-identification&lt;/a&gt;.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2017:news/2017-02/3</id>
    <title>Check out our new deep learning course ELEG 5491.</title>
    <link href="http://dl.ee.cuhk.edu.hk/"/>
    <updated>2017-02-01T00:00:00Z</updated>
    <content type="html">Check out our new deep learning course &lt;a href=&quot;http://dl.ee.cuhk.edu.hk/&quot;&gt;ELEG 5491&lt;/a&gt;.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2016:news/2016-10/2</id>
    <title>I received the ECCV 2016 Outstanding Reviewer Award.</title>
    <link href="https://xiaotong.me/news/"/>
    <updated>2016-10-01T00:00:00Z</updated>
    <content type="html">I received the ECCV 2016 Outstanding Reviewer Award.</content>
  </entry>
  <entry>
    <id>tag:xiaotong.me,2016:news/2016-09/1</id>
    <title>We were ILSVRC 2016 champions, winning 1st places in ImageNet detection, ImageNet video object tracking, and Places2 scene parsing!</title>
    <link href="http://image-net.org/challenges/LSVRC/2016/results"/>
    <updated>2016-09-01T00:00:00Z</updated>
    <content type="html">We were ILSVRC 2016 champions, winning &lt;a href=&quot;http://image-net.org/challenges/LSVRC/2016/results&quot;&gt;1st places&lt;/a&gt; in ImageNet detection, ImageNet video object tracking, and Places2 scene parsing!</content>
  </entry>
</feed>
