Siqi Yang

AI Research Scientist
Tencent
siqi.yang@uq.net.au
[Google Scholar] [CV (updated on Aug 2023)]

About

I am now an AI Research Scientist at Innovation and Entertainment Group, Tencent. My current works focus on the computer vision algorithms for video games and metaverse, such as text conditioned human motion generation, dance generation and game-character auto-creation. Prior to this, I was a Research Scientist for Game AI Team at Bytedance.

I received my Ph.D from the University of Queensland, Australia in 2020. I was lucky to have Prof. Brian C. Lovell and Dr. Arnold Wiliem as my advisors. I aslo collaborated with Dr. Jun Zhang, Dr. Xiao Han at Tencent AI Lab, Shenzhen. During my Ph.D. I was working in the robustness of computer vision applications, including domain adaptation, adversarial attacks and false positives.

I received my bachelor's degree from the South China University of Technology in 2015. During my undergraduate study, I spent one year as an exchange student at the University of Queensland and working on my thesis project, Advanced FPGA design with ZedBoard, with Dr. Adam Postula.

Work

  • 2022.04 - Present, Research Scientist, Tencent

  • 2020.11 - 2022.04, Research Scientist, Bytedance

Education

Doctor of Philosophy
Computer Science
Advisors: Prof. Brian Lovell and Dr. Arnold Wiliem
The University of Queensland (UQ), Australia
2016.01 - 2020.08

Bachelor of Engineering
Information Engineering
South China University of Technology (SCUT), China
2011.09 - 2015.07

Teaching

  • Teaching Assistant, Computer Vision & Image Processing (ELEC 4630), The Uni of QLD, 2017

  • Teaching Assistant, Digital Signal Processing (ELEC 4620), The Uni of QLD, 2017

Professional Activities

  • Reviewer, Pattern Recognition

  • Reviewer, Pattern Recognition Letters

  • Reviewer, IEEE Winter Conference on Applications of Computer Vision (WACV) 2017

Awards

  • Rank 3rd in the detection track of Visual Domain Adaptation Challenge at ECCV, 2018

  • UQ Graduate School Scholarship, 2017

  • 1st Place Presentation, in the research theme: Information Engineering at the EAIT Postgraduate Conference, The Uni of QLD, 2017

  • Third Prize, University Level, Mathematical Contest in Modeling, SCUT, 2013

  • Second Class Scholarship, SCUT, 2011, 2012

  • Excellent Student Learder, SCUT, 2012, 2013

Patents

Publications

LongDanceDiff: Long-term Dance Generation with Conditional Diffusion Model

Dance Generation Conditional Diffusion Model Generative AI Game and Virtual Huamn

Siqi Yang, Zejun Yang, Zhisheng Wang

arxiv, 2023

Abstract PDF

Regressive Domain Adaptation for 3D Game Character Auto-Creation using Synthetic Data

3D Game Character Creation Unsupervised Domain Adaptation Generative AI Game and Virtual Huamn

Siqi Yang, Feng Xiong, Yongqiang Gao, Yongqiang Mou, Hongliang Li

arxiv, 2022

Abstract

Minimizing Labeling Cost for Nuclei Instance Segmentation and Classification with Cross-domain Images andWeak Labels

Unsupervised Domain Adaptation Weakly-supervised Domain Adaptation Medical Image Processing Nuclei Instance Segmentation and Classification

Siqi Yang*, Jun Zhang*, Junzhou Huang, Brian C. Lovell, Xiao Han (*Co-fisrt Author)

AAAI Conference on Artificial Intelligence (AAAI), 2021

Abstract PDF

Unsupervised Domain Adaptive Object Detection using Forward-Backward Cyclic Adaptation

Unsupervised Domain Adaptation Meta Learning Object Detection

Siqi Yang, Lin Wu, Arnold Wiliem, Brian C. Lovell

Asian Conference on Computer Vision (ACCV), 2020

Abstract PDF Video

Using LIP to Gloss Over Faces in Single-Stage Face Detection Networks

Adversarial Attack Face Detection Object Detection

Siqi Yang, Arnold Wiliem, Shaokang Chen, Brian C. Lovell

European Conference on Computer Vision (ECCV) 2018

Abstract PDF Poster Demo video

It takes two to tangle: Cascading off-the-shelf face detectors

Face Detection False Positive Analysis

Siqi Yang, Arnold Wiliem, Brian C. Lovell

Computer Vision and Pattern Recognition(CVPR) Workshop on Biometrics, 2018

Abstract PDF Poster

TV-GAN: Generative Adversarial Network based Thermal to Visible Face Recognition

Face Recognition Generative Adversarial Network

Teng Zhang, Siqi Yang, Arnold Wiliem, Brian C. Lovell

International Conference on Biometrics (ICB), 2018

Abstract PDF

To Face or Not To Face: Towards Reducing False Positive of Face Detection

Face Detection False Positive Analysis

Siqi Yang, Arnold Wiliem, Brian C. Lovell

Image and Vision Computing New Zealand (IVCNZ), 2016

Abstract PDF

The GIST of Aligning Faces

Face Alignment

Siqi Yang, Arnold Wiliem, Brian C. Lovell

International Conference on Pattern Recognition (ICPR), 2016

Abstract PDF

Landmark manifold: Revisiting the Riemannian manifold approach for facial emotion recognition

Facial Emotion Recognition

Kun Zhao, Siqi Yang, Arnold Wiliem, Brian C. Lovell

International Conference on Pattern Recognition (ICPR), 2016

Abstract PDF

More About Me

I like playing badminton, dancing and photograhing.

During my undergraduate study, I was the Head of Social Event Team at the Student Union of School of Electronic & Information Engineering, SCUT, 2011-2013. We organzied the welcome events and annual ball for more than 500 students every year.

In the summer of 2012, I was lucky to be a short-term volunteer teacher in a rural region of China. My role was teaching English and dancing.