update: add OmniVTLA/RARK/Janus to research, refresh author bio and meta description

Author: cidxb

_config.yml

@@ -25,7 +25,7 @@ title:                 "Sid Liang, Yi"
 # A short description (~150 chars) of the page used for the meta description tag.
 # Can use markdown, but no more than one paragraph (enforced by `>`)
 description:           >
-  Personal website of Sid Liang, Yi.
+  Cognitive Robotics · Embodied AI — building robots that perceive and act like humans.
 
 # A shorter description for the sidebar.
 tagline:               Create a human-friendly future
@@ -46,9 +46,9 @@ author:
   name: Sid Liang, Yi
   email: cidxbx@gmail.com
   about: |
-    I am currently a Reinforcement Learning engineer at a Tactile sensor/ Dexterous hand / Humanoid robot startup.
-    
-    I got my bachelor's degree in Photoeletronic Science and Engineering from the Southwest University of Science and Technology, Sichuan, China.
+    Research Lead at Paxini Tech, building cognitive robots with tactile perception and multi-modal learning.
+
+    M.Sc. Sensor & Cognitive Psychology, TU Chemnitz. B.Sc. Photoelectronic Science, SWUST.
   
   picture:
     path: /assets/img/avatar.jpg

index.html

@@ -21,6 +21,7 @@
 <section id="news">
   <h2>News</h2>
   <ul>
+    <li><strong>[Aug, 2025]</strong> <a href="https://arxiv.org/abs/2508.08706" target="_blank">OmniVTLA</a> paper published — Vision-Tactile-Language-Action model achieving 100% success rate on dexterous manipulation tasks.</li>
     <li><strong>[Jan, 2026]</strong> Showcased embodied AI teleoperation system at CES 2026, Las Vegas.</li>
     <li><strong>[Nov, 2025]</strong> Demonstrated dexterous manipulation at IREX 2025, Tokyo.</li>
     <li><strong>[Mar, 2025]</strong> Exhibited at MWC 2025, Barcelona.</li>
@@ -36,6 +37,60 @@ <h2>News</h2>
 <section id="research">
   <h2>Research & Projects</h2>
 
+  <!-- OmniVTLA -->
+  <div class="publication">
+    <div class="publication-content">
+      <h3>OmniVTLA: Vision-Tactile-Language-Action Model with Semantic-Aligned Tactile Sensing</h3>
+      <p class="authors">
+        PaXini &amp; SJTU
+      </p>
+      <p class="venue">
+        <em>arXiv</em>, 2025
+      </p>
+      <p class="links">
+        <a href="https://arxiv.org/abs/2508.08706">paper</a>
+      </p>
+      <p class="description">
+        A multi-modal VLA model that fuses visual and tactile perception for dexterous manipulation. Achieves 100% success on contact-dense tasks with four-finger dexterous hands, significantly outperforming vision-only baselines.
+      </p>
+    </div>
+  </div>
+
+  <!-- RARK -->
+  <div class="publication">
+    <div class="publication-content">
+      <h3>RARK: Robot Agent Runtime Kernel</h3>
+      <p class="authors">
+        <strong>Yi Liang</strong>
+      </p>
+      <p class="venue">
+        <em>Open Source Project</em>, 2025
+      </p>
+      <p class="links">
+        <a href="https://github.com/cidxb/robot-agent-runtime-kernel">code</a>
+      </p>
+      <p class="description">
+        A lightweight runtime kernel for orchestrating robot agent behaviors, bridging perception, planning, and action execution in embodied AI systems.
+      </p>
+    </div>
+  </div>
+
+  <!-- Janus -->
+  <div class="publication">
+    <div class="publication-content">
+      <h3>Janus: Embodied AI Learning Framework</h3>
+      <p class="authors">
+        <strong>Yi Liang</strong> et al.
+      </p>
+      <p class="venue">
+        <em>Internal Open-Source Framework</em>, 2024–present
+      </p>
+      <p class="description">
+        An embodied AI learning framework built on NVIDIA Isaac Sim, adopted by 30+ developers for sim-to-real robot training, evaluation, and deployment.
+      </p>
+    </div>
+  </div>
+
   <!-- 主要研究项目 -->
   <div class="publication">
     <div class="publication-teaser">
@@ -118,7 +173,7 @@ <h3>Research Interests</h3>
 
   <p><strong>Embodied Cognition &times; Robot Learning</strong>: incorporating cognitive science findings and biological principles into robot learning algorithms; pursuing the fundamental question of how robots can build internal models that bridge raw sensory experience and meaningful action.</p>
 
-  <p>I have also been working on: Machine Learning of Graphs/Graph Theory, Vector Symbolic Architectures, Neuro-Symbolic AI, Causality Inference, and Fuzzy Logic.</p>
+  <p>I also maintain interest in neuro-symbolic AI, vector symbolic architectures, and causality inference — areas that inform my broader pursuit of brain-inspired machine intelligence.</p>
 
 </section>
 <section id="Me">