<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Manipulation |</title><link>https://caiyf03.github.io/tags/manipulation/</link><atom:link href="https://caiyf03.github.io/tags/manipulation/index.xml" rel="self" type="application/rss+xml"/><description>Manipulation</description><generator>HugoBlox Kit (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Mon, 01 Sep 2025 00:00:00 +0000</lastBuildDate><image><url>https://caiyf03.github.io/media/icon_hu_702a800cd775dbac.png</url><title>Manipulation</title><link>https://caiyf03.github.io/tags/manipulation/</link></image><item><title>GRASP Lab, University of Pennsylvania</title><link>https://caiyf03.github.io/internships/bin/</link><pubDate>Mon, 01 Sep 2025 00:00:00 +0000</pubDate><guid>https://caiyf03.github.io/internships/bin/</guid><description>&lt;h2 id="overview"&gt;Overview&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Graduate Researcher&lt;/strong&gt;
&lt;em&gt;University of Pennsylvania, 2025 – 2026&lt;/em&gt;&lt;/p&gt;
&lt;h3 id="world-models-for-robotic-manipulation"&gt;World Models for Robotic Manipulation&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;Conducted research on 3D scene reconstruction, temporal representation learning, and world models for embodied decision-making systems.&lt;/li&gt;
&lt;li&gt;Explored unified frameworks connecting perception, dynamic scene understanding, and planning for robotic manipulation tasks.&lt;/li&gt;
&lt;li&gt;Focused on learning environment representations that support downstream control and long-horizon reasoning.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id="training-and-evaluation-infrastructure"&gt;Training and Evaluation Infrastructure&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;Responsible for model training, experimental debugging, and large-scale dataset preprocessing.&lt;/li&gt;
&lt;li&gt;Optimized training workflows and experiment pipelines to improve reproducibility and computational efficiency.&lt;/li&gt;
&lt;li&gt;Participated in simulation environment design and benchmark evaluation for manipulation tasks.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3 id="research-analysis"&gt;Research Analysis&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;Surveyed and analyzed recent literature in embodied AI, generative world models, and robotic learning.&lt;/li&gt;
&lt;li&gt;Developed experimental comparison frameworks and performance evaluation protocols for research studies.&lt;/li&gt;
&lt;/ul&gt;
&lt;hr&gt;</description></item></channel></rss>