Proxy-less Architecture Search via Binarized Path Learning.
In ICLR, 2019

Sparsely Aggregated Convolutional Networks.
In ECCV, 2018

PDF Github

Learning to Forecast Videos of Human Activity with Multi-granularity Models and Adaptive Rendering.
In BMVC, 2018


Colorizing Color Images.
In HEVI, 2018

PDF Code

Attribute Recognition from Adaptive Parts.
In BMVC, 2016



[Aug 2018 ~ Currently] Visiting Student, HanLab @ MIT.

[Jan 2017 ~ Aug 2018] Research Intern, Sensetime @ Beijing.

[May 2017 ~ May 2018] Research Assistant, CVL @ SFU.

[May 2017 ~ Aug 2017] Deep Learning Engineer, TuSimple @ California.

[Sep 2015 ~ May 2017] Research Assistant, GruVi @ SFU.

Recent Talks

Sparsely Aggregated Convolution Networks
Jan 18, 2018
SparseNet: a new design pattern based on aggregation view
Nov 18, 2017
Deep Learning Tutorial for Computer Visioners
Oct 8, 2017
Neural Style Transform Explained
Apr 15, 2017
Deep Learning Tutorial for Beginners
Dec 8, 2016




A simple and flexible toolbox for MXNet framework.

Neural Machine Translation

Seq2seq + attention model for machine translation.

Machine Learning for Image Colorization

God says, let there be color, then the deep neural network brings the color.

Fast and Coherent Neural Style Transform for Videos

A new method for video stylization that keeps the temporal consistency and reasonable time costs