Yadong Mu

National Junior 1000-Talent Plan
Head of Machine Intelligence Lab
Institute of Computer Science and Technology
No. 128, Zhong-Guan-Cun North Street
Peking University, Beijing 100080, China
E-mail: myd AT pku.edu.cn OR muyadong AT gmail DOT com


I am leading the Machine Intelligence (MI) Lab at Institute of Computer Science & Technology, Peking Univeristy. Before joining Peking University since June 2016, I have ever worked as research fellow at National University of Singapore (PI: Prof. Shuicheng Yan), research scientist at the DVMM lab of Columbia University (PI: Prof. Shih-Fu Chang), researcher at the data mining team of Huawei Noah's Ark Lab in Hong Kong (head: Dr. Wei FAN), and senior scientist at Multimedia Department of AT&T Labs, New Jersey, U.S.A. (head: Dr. Behzad Shahraray). I obtained both the B.S. and Ph.D. degrees from Peking University.

I have interest in broad research topics in computer vision and machine learning, particularly large-scale video computing (search, indexing,  classification, event localization etc), human-centric visual analysis (pose estimation etc.), autonomous driving techniques, distributed / approximate large-scale machine learning, deep / reinforcement learning.

Research Directions

Video Semantic Understanding

Visual Search for
Gigantic Data


Machine Learning & Deep Learning Algorithm

Data Mining (e.g., on Telecom/Financial Data)

Selected Recent Research Projects

  • Autonomous Vehicle Steering Using Temporal Modeling
  • High-Capacity Convolutional Video Steganography
  • Crowd Counting and Localization in Survellience Videos
  • Large-Scale Video Classification with Multi-Modality Attentions
  • Proposal Forging Network in Video Action Localization


  • 本实验室招收意向保研北大计算机科学或数据科学方向的本科大三同学,另有招收多名大二本科实习生的计划,研究方向为计算机视觉和机器学习。欢迎有兴趣的同学联系(Email: myd@pku.edu.cn)!这里是招生说明。(12/2018)
  • In the spring semester of 2018, I will teach a new course "Computer Vision and Deep Learning" for undergraduate students in EECS, Peking University. (11/2017)
  • One paper collaborated with UESTC and Tecent AI Lab won the Best Paper Honourable Mentions at SIGIR 2017. [Link] (08/2017)
  • Call for Paper -- ACM Multimedia Workshop on Visual Analysis for Smart and Connected Communities. [Link] (06/2017)
  • I will teach a course advanced topics in computer vision" (course ID: 04802034), and co-teach the other course "deep learning" (course ID: 08408005) in the spring semester. The former will majorly discuss recent advances in computer vision and the latter will cover both deep learning theory and applications. (02/2017)
  • We won the first place out of 100+ teams in the "traffic sign detection in autonomous driving" competition (preliminary round) organized by China Computer Federation (CCF) and UISEE (a self-driving car startup). (11/2016)
  • Our team participated 2016 TRECVID MED (multimedia event detection) competition organized by National Institute of Standards and Technology (NIST). Our multi-modal MED system achieved top performance in three sub-tasks in the PS-100Ex setting. (10/2016)
  • Our team won the second place in RACV 2016 Iqiyi Video Annotation Challenge.
  • I am the recipient of National Junior 1000-Talent Plan and have joined the Institute of Computer Science and Technology, Peking University as a tenure track faculty and principal investigator. (05/2016)

Recent Publications [all]

  • Lu Chi, Yadong Mu: Learning End-to-End Autonomous Steering Model from Spatial and Temporal Visual Cues, ACM Multimedia Workshops 2017. [PDF]
  • Yadong Mu, Zhu Liu, Deep Hashing: A Joint Approach for Image Signature Learning, Thirty-First AAAI Conference (AAAI), 2017 [PDF]
  • Yadong Mu, Wei Liu, Cheng Deng, Zongting Lv, Xinbo Gao, Coordinate Discrete Optimization for Efficient Cross-View Image Retrieval, International Joint Conference on Artificial Intelligence (IJCAI) 2016 [PDF]
  • Yadong Mu, Fixed-Rank Supervised Metric Learning on Riemannian Manifold, Thirtieth AAAI Conference (AAAI), 2016 [PDF] (code by request)
  • Yadong Mu, Wei Liu, Xiaobai Liu, Wei Fan, Stochastic Gradient Made Stable: A Manifold Propagation Approach for Large-Scale Optimization , to appear in IEEE Transactions on Knowledge and Data Engineering (TKDE), 2016 [PDF]


Thanks the generous support of all the sponsors.