About Me

I am a research engineer in Google DeepMind, applying DeepMind's cutting edge research work onto Google's large-scale applications. I work closely with Alphabet partners such as Google Research, YouTube and Google Ads. I am broadly interested in fundamental research and scalable applications on Computer Vision and Natural Language Processing.

Experience

07/2020 - Present Research Engineer, DeepMind

05/2019 - 09/2019 Research Intern, Facebook AI

05/2018 - 11/2018 Research Intern, Nvidia Research

06/2017 - 04/2018 Research Intern, Facebook AI Research (FAIR)

Projects

• Auto-generated Chapters, applied to 80M YoutTube videos: covered in Google I/O feature (10:53)

• YouTube Video Content Recognition / Temporal Localization

• Visual Relationship Detection: 1st place in the OpenImages Relationship Detection Challenge

• Grounded Visual Question Answering: 2nd place in the Stanford GQA Challenge

Publications

2nd Place Solution to the GQA Challenge 2019

Shijie Geng*, Ji Zhang*, Hang Zhang, Ahmed Elgammal, Dimitris N. Metaxas

CVPR 2019 Workshop

paper bibtex

Graphical Contrastive Losses for Scene Graph Parsing

Ji Zhang, Kevin J. Shih, Ahmed Elgammal, Andrew Tao, Bryan Catanzaro

CVPR 2019

paper bibtex code

Large-Scale Visual Relationship Understanding

Ji Zhang, Yannis Kalantidis, Marcus Rohrbach, Manohar Paluri, Ahmed Elgammal, Mohamed Elhoseiny

AAAI 2019 (Spotlight)

paper bibtex code slides

An Interpretable Model for Scene Graph Generation

Ji Zhang, Kevin Shih, Andrew Tao, Bryan Catanzaro, Ahmed Elgammal

NeurIPS 2018 workshop

paper bibtex slides

Relationship Proposal Networks

Ji Zhang, Mohamed Elhoseiny, Scott Cohen, Walter Chang, Ahmed Elgammal

CVPR 2017

paper bibtex

Visual Tracking via Supervised Similarity Matching

Ji Zhang, Jie Sheng, Ankur Teredesai

ACCV 2014

paper bibtex