Jian Yao

Jian's Picture

Contact Information:

Address:
Toyota Technological Institute at Chicago
6045 S. Kenwood Ave.
Chicago, IL 60637
Phone: 773-834-2556
Email: yaojian@{ttic.edu | uchicago.edu}
     

News:

From January 2014, I will join in Machine Learning Group at University of Toronto to continue with my PhD (expecting to graduate in the summer, 2015).
This website will not be updated, please refer to my new Personal Website

Short Biography:

Since September 2010, I have been a PhD student working with Professon Raquel Urtasun in Toyota Technological Institute at Chicago (TTIC),
a philanthropically endowed academic computer science institute located on University of Chicago campus.
I studied as a Master Student in the Image Processing and Pattern Recognition Lab(2009-2010), Shanghai JiaoTong University, Shanghai, China.
I received the Bachelor Degree on Electrical Engineering(2005-2009) in Xi'an JiaoTong University, Xi'an, China.
I am interested in the following areas:
  • Computer Vision
  • Machine Learning
  • Optimization

Publications

  • R. Mottaghi, S. Fidler, J. Yao, R. Urtasun and D. Parikh
    Analyzing Semantic Segmentation Using Human-Machine Hybrid CRFs [PDF]
    In Conference of Computer Vision and Pattern Recognition (CVPR), 2013

  • Jian Yao, Sanja Fidler and Raquel Urtasun
    Describing the Scene as a Whole: Joint Object Detection, Scene Classification and Semantic Segmentation [PDF]
    In Conference of Computer Vision and Pattern Recognition (CVPR), 2012

Projects

  • Holistic Scene Understanding

  • In this project, we propose an approach to holistic scene understanding that jointly reasons about regions,
    location, class and spactial extent of objects, presence of a class in the images, as well as the scene type.

    In the model of Markov Random Field which relates all the information together, learning and inference
    are efficient as we build the graphical model at the segment level, and introduce auxiliary variables that
    allow us to decompose the inherent high-order potentials into pairwise potentials between a few variables
    with small number of states.

    Finally, this model not only improves segmentation but also detection as well as scene classification.

Teaching

  • University of Chicago, CSPP 55001: Algorithms, Autumn 2011. TA, Instructor: Gerry Brady
  • Shanghai JiaoTong University, 2010 Spring, Laboratory in Electrics and Electronics, TA

Professional Activities

  • Internship at Mitsubishi Electric Research Lab (MERL), working with Srikumar Ramalingam, Summer 2013
  • Research Internship at Rice University with Prof. Ashok Veeraraghavan, Summer 2012
    Depth Estimation based on the Light Field Camera
  • The organizer of TTIC Vision Reading Group
    Vision Reading Group at TTIC
  • The active member in Bookworm Club for Vision
    Book Club for Vision
  • Volunteer at In Conference of Computer Vision and Pattern Recognition (CVPR) 2012

Selected Graduate Courses

  • Machine Learning
  • Visual Recognition
  • Computer Vision
  • Probabilistic Graphical Model
  • Linear Programming
  • Algorithm

Personal interests

  • Tennis and Soccer
  • Now I find much fun in skiing and skating in Chicago's Winter