Mark Dong

Jin-Dong Dong (Mark)

Computer Vision / Machine Learning Learner

Contact Me

About Me

I will be joining Jia-Bin Huang's lab in late November as a visiting student. In the summer, I did my internship at DeePhi Tech (Xilinx) where I worked on Multi-Task Architecture Search. Previously spent a wonderful year as a research assistant at NTHU VSLab under the supervision of prof. Min Sun.

My recent works include device-aware architecture search and predicting saliency map in 360o videos, I'm also interested in meta-learning, vision-language and general video understanding problems. Hobbies are playing music and sports!

Experiences (More information here)

Intern - DeePhi Tech (Xilinx) (Summer, 2018)

Research Assistant - VSLab, National Tsing-Hua University (June, 2017 - June, 2018)

Advisor: Prof. Min Sun

Intern - Multimedia Lab, Institute of Information Science, Academia Sinica (Summer, 2017)

Advisor: Dr. Tyng-Luh, Liu

Publications

DPP-Net: Device-aware Progressive Search for Pareto-optimal Neural Architectures

J.-D. Dong, A.-C. Cheng, D.-C. Juan, W. Wei, M. Sun. ECCV 2018, ICLR 2018 Workshop

We propose a multi-objective architectural search framework to automatically generate networks that achieve Pareto Optimality.

Cube Padding for Weakly-Supervised Saliency Prediction in 360° Videos

H.-T. Zheng, C.-H. Chao, J.-D. Dong, H.-K. Wen, T.-L. Liu, M. Sun. CVPR 2018

We propose a spatial-temporal network which is (1) weakly-supervisedly trained and (2) tailor-made for 360o viewing sphere, to predict saliency map for 360o video.

My GitHub