↓Skip to main content

PKU ZhiClass

PKU ZhiClass

———Assembly of Intellect———

Welcome to Zhi-Class

The Experimental Class in Intelligence Science and Technology, commonly known as "Zhi-Class," is a collaborative initiative by the School of Intelligence Science and Technology and the School of Electronics Engineering and Computer Science at Peking University. Building on the legacy of China's first educational program in Intelligence Science and Technology, Zhi-Class emphasizes a strong foundation in mathematics and physics, deep interdisciplinary integration of science and engineering, systematic curriculum planning, and internationally renowned faculty. The program is dedicated to cultivating the next generation of leaders in AI for China. For detailed introduction and contact information, please visit the Intro page.


Recent
#

2026毕业特辑-徐佳怡 | 风物长宜放眼量 ↗ ↖
PKU智班
活动
2022级智班毕业专访——徐佳怡学姐篇
2026毕业特辑-戴冠哲 | 破而后立,在平淡与抗挫中寻找生活的解 ↗ ↖
PKU智班
活动
2022级智班毕业专访——戴冠哲学长篇
2026毕业特辑-陈灵钧 | You will never walk alone. ↗ ↖
PKU智班
活动
2022级智班毕业专访——陈灵钧学长篇
LDA-1B: Scaling Latent Dynamics Action Model via Universal Embodied Data Ingestion ↗ ↖
Jiangran Lyu*
, 
Kai Liu*
, 
Xuheng Zhang*
, 
Haoran Liao
, 
Yusen Feng
, 
Wenxuan Zhu
, 
Tingrui Shen
, 
Jiayi Chen
, 
Jiazhao Zhang
, 
Yifei Dong
, 
Wenbo Cui
, 
Senmao Qi
, 
Shuo Wang
, 
Yixin Zheng
, 
Mi Yan
, 
Xuesong Shi
, 
Haoran Li
, 
Dongbin Zhao
, 
Ming-Yu Liu
, 
Zhizheng Zhang
, 
Li Yi
, 
Yizhou Wang
, 
He Wang
Arxiv Github RSS 在投
Recent robot foundation models largely rely on large-scale behavior cloning, which imitates expert …
NavSpace: How Navigation Agents Follow Spatial Intelligence Instructions ↗ ↖
Haolin Yang*
, 
Yuxing Long*
, 
Zhuoyuan Yu
, 
Zihan Yang
, 
Minghan Wang
, 
Jiapeng Xu
, 
Yihan Wang
, 
Ziyan Yu
, 
Wenzhe Cai
, 
Lei Kang
, 
Hao Dong1,2,‡
Arxiv ICRA
Instruction-following navigation is a key step toward embodied intelligence.
Learning Physics-Grounded 4D Dynamics with Neural Gaussian Force Fields ↗ ↖
Shiqian Li
, 
Ruihong Shen
, 
Junfeng Ni
, 
Chang Pan
, 
Chi Zhang
, 
Yixin Zhu
Arxiv Github ICLR
Predicting physical dynamics from visual data remains a fundamental challenge in AI, as it requires …