Junyuan "Eric" Xie (解浚源)

Ph.D Student @ Grail Lab, CSE, Univeristy of Washington
jxie <at> cs.washington.edu

185 Stevens Way, CSE 506
Paul G. Allen Center, Box 352350
Seattle, WA 98195-2350
CV  |  Google Scholar  |  Github

About me

I'm a 3rd year Ph.D student studying computer science at University of Washington working with Ali Farhadi and Ross Girshick. Before coming to UW, I received my Bachelor's degree in CS from University of Science and Technology of China.

My research interests lie in machine learning and computer vision. In particular, I'm interested in discovering new ways of training deep neural networks with unlabeled or "naturally" labeled data that's easy to obtain.

I also like to work on open source software projects for research and fun.

Software

MXNet

MXNet is an open source deep learning framework designed for both efficiency and flexibility. With MXNet, it is easy to experiment with new models as well as scaling them to multiple GPUs or clusters.
I am one of the main contributors to this project.
Github  

Publications

Deep3D: Fully Automatic 2D-to-3D Video Conversion with Deep Convolutional Neural Networks

Junyuan Xie, Ross Girshick, Ali Farhadi
European Conference on Computer Vision (ECCV), 2016. [PDF] [Code]

Unsupervised Deep Embedding for Clustering Analysis

Junyuan Xie, Ross Girshick, Ali Farhadi
International Conference on Machine Learning (ICML), 2016. [PDF]

Image Denoising and Inpainting with Deep Neural Networks

Junyuan Xie, Linli Xu, Enhong Chen
Advances in Neural Information Processing Systems (NIPS), 2012. [PDF]

Experience

Software Engineering Intern, Ph.D

Google  |  Summer, 2015

Research Intern

Microsoft Research Asia  |  2012 - 2013