Home People Research Publications Demos
News Jobs Prospective
About Internal

Weixin Li Weixin Li (李伟欣)
EBU-1, Room 4604
9500 Gilman Drive
La Jolla, CA 92093-0409

Brief Biography

I physically logged in this world at Guangxi, an autonomous region full of spectacular mountain scenery in southern China, and spent the whole childhood there until high school. I left hometown for undergraduate education at Tsinghua University, Beijing in 2004, where I received my bachelor's degree from Department of Automation, School of Information Science and Technology, at the end of the marvelous college life in 2008. Shortly after that, I travelled across the Pacific Ocean to the United States of America, in pursuit of graduate study at the gorgeous city of San Diego. After years of sun-and-beach obsession, I obtained my Ph.D. degree from UC San Diego, under the supervision of Professor Nuno Vasconcelos at Statistical Visual Computing Lab in the Department of Electrical and Computer Engineering. During this quest, I was very fortunate to be guided, get enlightened and collaborate with many visionary minds with inspiration and insights from both academia and industry, and felt deeply honored to have received various forms of recognition and supports from multiple sources. In the next chapter of the odyssey, I set off on a new journey onboard the artificial intelligence program of Facebook as a research scientist, looking forward to making impacts to billions of people worldwide via the cybernetic connections.


Tsinghua University
B.E. in Automatic Control with honors
Beijing, China
Admitted via Chinese Science Olympiads with the National College Entrance Exam waived
University of California, San Diego
M.Sc. in Electrical Engineering
La Jolla, CA, United States
Thesis: Anomaly Detection and Localization in Crowded Scenes
University of California, San Diego
Ph.D. in Electrical Engineering
La Jolla, CA, United States
Dissertation: Visual Understanding of Complex Human Behavior via Attribute Dynamics


My research interest and curiosity mainly reside in computational vision and machine learning, which aim to answer two fundamental questions in the vision science via the computating machinery: what are the principled ways to represent the huge variability of visual phenomena around us; and how to recover the latent governing factors, their relationships, and the associated structures underneath that produce such a magnificent visual world. Specifically, I have been focusing on statistical treatments for (1) discriminative and generative models with latent variables (e.g., dynamical systems, recurrent architectures); (2) sequential signal modeling and prediction; and (3) human behavior, activity and event analysis in video. Most of these research works have been documented in the format of peer-reviewed academic publications. Most of my research works at graduate school were supported by grants from the National Foundation of Science (NSF) .


My online publication profile can be found at Google Scholar, Research Gate, DBLP, or Semantic Scholar. (Note: given the overwhelming ambiguity of Chinese names when romanized, I may show up as "Wei-Xin LI", i.e., with surname all capitalized and given name hyphenated, in publications for better identifiability.)

Complex Activity Recognition via Attribute Dynamics
Wei-Xin LI and Nuno Vasconcelos
International Journal of Computer Vision (IJCV)
Vol. 122 (2), pp. 334-370, April 2017
[ project | preprint (pdf) | DOI | BibTeX ]

VLAD3: Encoding Dynamics of Deep Features for Action Recognition
Yingwei Li, Wei-Xin LI, Vijay Mahadevan and Nuno Vasconcelos
Proc. of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR)
Las Vegas, Nevada, United States, 2016
[ project | pdf | DOI | BibTeX ]

Multiple Instance Learning for Soft Bags via Top Instances
Wei-Xin LI and Nuno Vasconcelos
Proc. of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR)
Boston, Massachusetts, United States, 2015
[ project | pdf | DOI | BibTeX ]

Anomaly Detection and Localization in Crowded Scenes
Wei-Xin LI, Vijay Mahadevan and Nuno Vasconcelos
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
Vol. 36 (1), pp. 18-32, January 2014
[ project | dataset | pdf | DOI | appendix ( pdf ) | BibTeX ]

Dynamic Pooling for Complex Event Recognition
Wei-Xin LI, Qian Yu, Ajay Divakaran and Nuno Vasconcelos
Proc. of IEEE International Conf. on Computer Vision (ICCV)
Sydney, New South Wales, Australia, 2013
[ project | pdf | DOI | BibTeX ]

Recognizing Activities via Bag of Words for Attribute Dynamics
Wei-Xin LI, Qian Yu, Harpreet Sawhney and Nuno Vasconcelos
Proc. of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR)
Portland, Oregon, United States, 2013
[ project | pdf | DOI | BibTeX ]

Recognizing Activities by Attribute Dynamics
Wei-Xin LI and Nuno Vasconcelos
Advances in Neural Information Processing Systems (NIPS)
Lake Tahoe, Nevada, United States, 2012
[ project | pdf | BibTeX ]

Anomaly Detection in Crowded Scenes
(oral presentation)
Vijay Mahadevan, Wei-Xin LI, Viral Bhalodia and Nuno Vasconcelos
Proc. of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR)
San Francisco, California, United States, 2010
[ project | dataset | pdf | DOI | BibTeX ]

Professional Experience

Research Scientist
Menlo Park, CA
United States
2016 - present
Statistical Visual Computing Lab
UC San Diego

Graduate Research/Teaching Assistant
La Jolla, CA
United States
2009 - 2016
Google Research
Research Intern
New York City
United States
SRI International
Research Intern
Princeton, NJ
United States
Microsoft Research Asia (MSRA)
Research Intern
Institute of Image and Graphics
Tsinghua University

Undergraduate Research Assistant

Major Awards and Honors

Chinese Government Award for Outstanding Students Abroad 2013
by Chinese Scholarship Council
Graduate Fellowship 2008
by UC San Diego
Outstanding College Graduate 2008
by Tsinghua University & Beijing
National Scholarship 2007
by Ministry of Education of China
Siemens Scholarship 2006
by Siemens AG
HSBC Scholarship 2005
by Hongkong Bank Foundation of HSBC

October 2016