2018.06 Taskonomy Received CVPR 2018 Best Paper Award!
2018.06 Received a B.S. in computer science with Departmental Honor and University Distinction
2018.03 Received Frederick Emmons Terman Engineering Scholastic Award (Top 5% of entire Stanford Engineering School)
2018.02 Paper accepted as Oral at CVPR2018 (Taskonomy: Disentangling Task Transfer Learning) [website]
2018.02 Admitted to Stanford University CS Ph.D.!
Sep. 2014 - Jun. 2018, Department of Computer Science, Stanford University,
Undergraduate Student. GPA: 4.01/4.00
Terman Award Winner, CS Department Honor, University Distinction
Advised by Prof. Silvio Savarese
Jun. 2016 - Sep. 2016, Project Fi, Google Inc.,
Real time phone call transcription service
Mentor: Madhu R. Adupala
Aug. 2016 - Sep. 2016, Google Brain, Google Inc.,
Measuring Gradient Descend Batch Variance
Mentor: Alex Davies
Amir R. Zamir, William B. Shen*, Alexander Sax*, Leonidas Guibas, Jitendra Malik, Silvio Savarese. Taskonomy: Disentangling Task Transfer Learning.
[Blog Post in Chinese (知乎)]
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Best Paper 2018
Research on visual task space’s structure; leveraging task space structure to optimize supervision policy of a set of tasks, the models learned using recommended transfers achieve much better performance than the models trained from scratch and come close to models that are trained with an order of magnitude more data with full-supervision.
Kuo-Hao Zeng, William B. Shen, De-An Huang, Min Sun, Juan Carlos Niebles. Visual Forecasting by Imitating Dynamics in Natural Sequences.
IEEE International Conference on Computer Vision (ICCV), Spotlight 2017.
Research on a general framework for visual forecasting, which directly imitates visual sequences by formulating visual forecasting as an inverse reinforcement learning (IRL) problem.
Research on novel feedback network paradigm that offers advantages including early prediction, taxonomic compliance and curriculum-based learning over traditional feedforward counterpart.
CS231A Course Project: William B. Shen, Song Han, Zuozhen Liu. Drone Human Tracking Using Faster RCNN and KCF.
Course project on implementing Faster-RCNN to detect human and KCF to track human on drones. Heavy optimization with frame-rate using TX1/TK1.
Stanford CS106A Graphics Contest Winner (Prof. Mehran Sahami, Autumn 2014)
Probably the nerdest thing I have done...
A weird mash-up of Mario, Galagal, Pacman, Star War, RPG.
© William Shen 2017