Research Scientist Google Brain Contact Google LLC 1600 Amphitheatre Pkwy Mountain View, CA 94043 USA

Biography

I am currently a research scientist at Google Brain. I am working on deep learning and its applications in text-to-speech and speech recognition.

I completed my PhD at MIT working with Prof. Vivienne Sze on energy efficient autonomous navigation algorithms and systems.

I complete my master thesis under the supervision of Prof. Edward H. Adelson on computational photography.

Before I joined MIT, I worked at Microsoft Research Asia with Prof. Yi Ma on 3D reconstruction and rank minimization.

I finished my undergraduate study at Tsinghua Xuetang Special Pilot CS Class (Yao Class) supervised by Turing Award Winner Prof. Andrew Chi-Chih Yao in Institute of Interdisciplinary Information Sciences, Tsinghua University.

My research interest spans the area of sparsity, low-rank matrix recovery, symmetry/regularity of textures, 3D computer vision, computational photography, object detection, video compression, localization, mapping.

You can find my resume here

Education

• 2014 - 2019, PhD in Electrical Engineering and Computer Science
  Massachusetts Institute of Technology
  Advisor Prof. Vivienne Sze

• 2012 - 2014, Master of Science in Computer Science
  Massachusetts Institute of Technology
  Advisor Prof. Edward H. Adelson

• 2007 - 2011, Bachelor of Engineering in Computer Science
  Tsinghua Xuetang Special Pilot CS Class (Yao Class), Tsinghua University
  Advisor Prof. Andrew Chi-Chih Yao

Experience

• 2018.2 - present, Research Scientist
  Google Brain, Mountain View, California

• 2016.6 - 2016.7, Visiting Researcher
  Shanghai Tech University, Shanghai, China

• 2015.12 - 2016.1, Visiting Researcher
  National Institute of Informatics, Tokyo, Japan

• 2014.5 - 2014.8, Research intern
  National Institute of Informatics, Tokyo, Japan

• 2013.5 - 2013.8, Software Engineer Intern
  Google Mountain View, California

• 2011.7 - 2012.6, Assistant Researcher
  Visual Computing group, Microsoft Research Asia

• 2009.10 - 2011.6, Research Intern
  Visual Computing Group, Microsoft Research Asia

Journal Papers

• FSMI: Fast computation of Shannon Mutual Information for information theoretic mapping
  Zhengdong Zhang, Trevor Henderson, Sertac Karaman, and Vivienne Sze
  submitted to International Journal of Robotics Research (IJRR), 2019

• Navion: A 2-mW Fully Integrated Real-Time Visual-Inertial Odometry Accelerator for Autonomous Navigation of Nano Drones
  Amr Suleiman, Zhengdong Zhang, Luca Carlone, Sertac Karamand, and Vivienne Sze
  IEEE Journal of Solid-State Circuits (JSSC), 2019

• A 58.6 mW 30 Frames/s Real-Time Programmable Multiobject Detection Accelerator With Deformable Parts Models on Full HD 1920x1080 Videos
  Amr Suleiman, Zhengdong Zhang, Vivienne Sze
  IEEE Journal of Solid-State Circuits (JSSC), 2017

• Texture Repairing by Unified Low Rank Optimization
  Xiao Liang, Xiang Ren, Zhengdong Zhang, Yi Ma
  Journal of Computer Science and Technology 31 (3), 525-546, 2016
  

• Robust Subspace Discovery via Relaxed Rank Minimization
  Xinggang Wang, Zhengdong Zhang, Yi Ma, Xiang Bai, Wenyu Liu, and Zhuowen Tu
  Neural Computation 2014, conference version presented at ACCV 2012 [pdf]

• TILT: Transform Invariant Low-rank Textures
  Zhengdong Zhang, Arvind Ganesh, Xiao Liang and Yi Ma
  International Journal of Computer Vision (IJCV) 2011, conference version presented at ACCV 2010 [pdf, project website]

Conference Papers

• High-Through Computation of Shannon Mutual Information on Chip
  Peter Li*, Zhengdong Zhang*, Sertac Karaman, and Vivienne Sze
  Robotics: Science and Systems (RSS) 2019

• FSMI: Fast computation of Shannon Mutual Information for information-theoretic mapping
  Zhengdong Zhang, Trevor Henderson, Vivienne Sze, and Sertac Karaman
  IEEE Conference on Robotics and Automation (ICRA) 2019

• Navion: A 2-mW Fully Integrated Energy-Efficient Visual-Inertial Odometry Accelerator for Autonomous Navigation of Nano Drones
  Amr Suleiman, Zhengdong Zhang, Luca Carlone, Sertac Karamand, and Vivienne Sze
  2016 Symposia on VLSI Technology and Circuits (VLSI) 2018

• Visual-Inertial Odometry on Chip: An Algorithm-and-Hardware Co-design Approach
  Zhengdong Zhang*, Amr Suleiman*, Luca Carlone, Vivienne Sze, Sertac Karaman
  Robotics: Science and Systems (RSS) 2017
  

• Hardware for Machine Learning: Challenges and Opportunities
  Vivienne Sze, Yu-Hsin Chen, Joel Emer, Amr Suleiman, Zhengdong Zhang
  arXiv preprint arXiv: 1612.07625
  

• FAST: Free Adaptive Super-Resolution via Transfer for Compressed Videos
  Zhengdong Zhang, Vivienne Sze
  CVPR Workshop 2017

• A 58.6mW Real-Time Programmable Object Detector with Multi-Scale Multi-Object Support Using Deformable  Parts Model on 1920x1080 Video at 30fps
  Amr Suleiman, Zhengdong Zhang, and Vivienne Sze
  2016 Symposia on VLSI Technology and Circuits (VLSI) 2016

• Rotate Intra Block Copy for Still Image Coding
  Zhengdong Zhang, and Vivienne Sze
  International Conference on Image Processing (ICIP) 2015

• Sparkle Vision: Seeing the World through Random Specular Microfacets
  Zhengdong Zhang, Phillip Isola, and Edward H. Adelson
  Workshop on Computational Camera and Display, IEEE Conference on Computer Vision and Pattern Recognition (CVPR-CCD) 2015

• Repairing Sparse Low-rank Textures
  Xiao Liang, Xiang Ren, Zhengdong Zhang, and Yi Ma
  European Conference on Computer Vision (ECCV) 2012[pdf]

• Rectifying Low-rank Textures on Generalized Cylindrical Surfaces
  Zhengdong Zhang, Xiao Liang, and Yi Ma
  International Conference on Computer Vision (ICCV) 2011[pdf]

• Camera Calibration with Lens Distortion from Low-rank Textures
  Zhengdong Zhang, Yasuyuki Matsushita, and Yi Ma
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2011[pdf]

• Decomposing Background Topics from Keywords by Principal Component Pursuit
  Kerui Min, Zhengdong Zhang, John Wright, and Yi Ma
  ACM International Conference on Information and Knowledge Management (CIKM) 2013[pdf]

Patent

• Robust Recovery of Transform Invariant Low-rank Textures
  Zhengdong Zhang, Xiao Liang, Arvind Ganesh, and Yi Ma
  U.S. Patent Application Serial No. 12/955,737, filed on November 29, 2010

• Robust Camera Calibration and Camera Undistortion
  Zhengdong Zhang, Yasuyuki Matsushita, and Yi Ma
  MS NO. 333143.01

• Rectification of Characters and Text as Transform Invariant Low-rank Textures
  Xin Zhang, Zhengdong Zhang, Xiao Liang, Zhouchen Lin, and Yi Ma
  MS No. 33145.01

• Leading Edge Detection, Rotation, and Occlusion of Space between Neighboring Image Pairs of a Set of Images
  Zhengdong Zhang, and Yuli Gao
  U.S. Patent Application No. 14/173,785, Google Ref. No. GP-19754-0-US

Classes at MIT

• 6.438 Algorithm for Inference

• 6.869 Advances in Computer Vision

• 6.875 Cryptography and Cryptanalysis

• 6.820 Foundation of Programming Analysis

• 6.344 Digital Image Processing (Teaching Assistant)

• 6.374 Analysis and Design: Digital Circuits

• 15.401 Finance Theory I

• 6.870 Smartphone Vision

• 15.439 Investment Management

Award and Honors

• David S. Y. Wong Fellowship at MIT, 2012

• Kaifeng Scholarship at Tsinghua University, 2009

• Freshman Scholarship at Tsinghua University, 2008

• Ranked 2/280758 in National College Entrance Examination, Sichuan Province, China, 2007

• First Prize Award in National Mathematics Contest, 2006