Qiang Meng (孟强)

I am a senior engineer DiDi focusing on autonomous driving. Previously, I was an algorithm engineer and a sub-core team leader at Aibee, where I worked on computer vision tasks including face recognition, image retrieval, car/person re-identification, etc.

I graduated with a master degree in Industrial Engineering from University of Washington, Seattle. Before that, I received my bachelor degree in Mechanical Engineering from University of Science and Technology of China, with a GPA of 3.78/4.3 (Rank: 3/61).

E-mail  |  Curriculum Vitae  |  Publications  |  Github


News
  • 08/2022:    Release the paper "Towards Privacy-Preserving, Real-Time and Lossless Feature Matching" on arXiv.
  • 04/2022:    Invited talk in Beijing Jiaotong University.
  • 03/2022:    Invited talk in ICLR 直播分享会 organized by ReadPaper.
  • 01/2022:    Our Paper "Improving Federated Learning Face Recognition via Privacy-Agnostic Clusters" was accepted by ICLR 2022 as a SPOTLIGHT paper.
  • 01/2022:    Release the paper "Basket-based Softmax" on arXiv.
  • 11/2021:    Invited talk in the workshop on face image quality organized by EAB, DHS-OBIM, NIST, eu-LISA, etc.
  • 07/2021:    Our paper "Learning Compatible Embeddings" was accepted by ICCV 2021.
  • 07/2021:    Release the paper "PoseFace: Pose-Invariant Features and Pose-Adaptive Loss for Face Recognition" on arXiv. .
  • 06/2021:    Invited talk about MagFace in CVPR 论文分享会 organized by 机器之心.
  • 03/2021:    Our paper "MagFace: A Universal Representation for Face Recognition and Quality Assessment" was accepted by CVPR 2021 as an ORAL paper.
  • 12/2020:    Our paper "Searching for Alignment in Face Recognition" was accepted by AAAI 2021.

First-author Publications
My research interests lie in computer vision, deep learning and optimization. Representative works are highlighted.

Towards Privacy-Preserving, Real-Time and Lossless Feature Matching
Qiang Meng, Feng Zhou
arXiv, 2022
paper | code
SecureVector is a plug-in module that achieves real-time and lossless feature matching among sanitized features, along with much higher security levels than state-of-the-arts.
Improving Federated Learning Face Recognition via Privacy-Agnostic Clusters
Qiang Meng, Feng Zhou, Hainan Ren, Tianshu Feng, Guochao Liu, Yuanqing Lin
ICLR , 2022 (Spotlight)
project page | paper | 知乎
A practical framework that significantly improves federated learning performance on face recognition with privacy guarantees. The keys include a well-designed differentially private local clustering mechanism and a consensus-aware recognition loss.
Learning Compatible Embeddings
Qiang Meng, Chixiang Zhang, Xiaqiang Xu, Feng Zhou
ICCV, 2021
project page | paper | 知乎 | short video | code
A general framework (LCE) which is applicable for both cross model compatibility and compatible training in direct/forward/backward manners.
MagFace: A Universal Representation for Face Recognition and Quality Assessment
Qiang Meng, Shichao Zhao, Zhida Huang, Feng Zhou
CVPR, 2021 (Oral presentation)
project page | paper | 知乎 | short video | code
A novel loss which equips feature magnitudes with the ability to represent face qualities, as well as achieves better performances on face recognition and clustering. More importantly, no additional labels are required!
Basket-based Softmax
Qiang Meng, Guxin Qian, Xiaqing Xu, Feng Zhou
arXiv, 2022
A simple but effective mining-during-traning strategy which trains models on multiple datasets in an end-to-end fashion.
PoseFace: Pose-Invariant Features and Pose-Adaptive Loss for Face Recognition
Qiang Meng, Xiaqing Xu, Xiaobo Wang, Yang Qian, Yunxiao Qin, Zezheng Wang, Chenxu Zhao, Feng Zhou, Zhen Lei
arXiv, 2021
An efficient large-pose face recognition method which utilizes the facial landmarks to disentangle the pose-invariant features and exploits a pose-adaptive loss to handle the imbalance issue adaptively.

Co-author Publications
Searching for Alignment in Face Recognition
Xiaqing Xu, Qiang Meng, Yunxiao Qin, Jianzhu Guo, Chenxu Zhao, Feng Zhou, Zhen Lei
AAAI, 2021
paper | 知乎
We design a face template searching space with decomposed crop size and vertical shift, and propose the Face Alignment Policy Search (FAPS) to find optimal alignment templates for face recognition.
CHI: A Contemporaneous Health Index for Degenerative Disease Monitoring using Longitudinal Measurements
Yijun Huang, Qiang Meng, Heather Evans, William Lober, Yu Cheng, Xiaoning Qian, Ji Liu, Shuai Huang
Journal of biomedical informatics, 2017
A optimization formulation is developed for contemporaneous patient risk monitoring by exploiting the emerging data-rich environment in healthcare applications.

Honors
  • [2015] College of Engineering Dean's Fellowship, University of Washington
  • [2014] Samsung Scholarship, University of Science and Technology of China
  • [2013] National Encouragement Scholarship, University of Science and Technology of China
  • [2013] First prize in the Challenge Cup, University of Science and Technology of China
  • [2012] National Encouragement Scholarship, University of Science and Technology of China
  • [2012] 1st place in RoboGame Robot Competition, University of Science and Technology of China

Teaching
  • IND E 315 - Probability and Statistics for Engineers [Spring 17] [Summer 17] [Winter 18] [Spring 18]
  • IND E 250 - Linear and Network Programming [Fall 16] [Fall 17]
  • IND E 410 - Fundamentals of Engineering Economy [Winter 16]


Website adapted from Jon Barron & Amlaan Bhoi Last updated in Aug 2022