AI models on Edge

Quantization and Training of Low Bit-Width Convolutional Neural Networks for Object Detection
with Penghang Yin, Yingyong Qi, Jack Xin.
Journal of Computational Mathematics, 2018
Featured article!

BinaryRelax: A Relaxation Approach For Training Deep Neural Networks With Quantized Weights
with Penghang Yin, Jiancheng Lyu, Stanley Osher, Yingyong Qi, Jack Xin.
SIAM Journal on Imaging Sciences, 2018
Code

Blended coarse gradient descent for full quantization of deep neural networks
with Penghang Yin, Jiancheng Lyu, Stanley Osher, Yingyong Qi, Jack Xin.
Research in the Mathematical Sciences, Springer, 2018
Code

DAC: Data-free Automatic Acceleration of Convolutional Networks
with Xin Li, Bolan Jiang, Yingyong Qi, Mooi Choo Chuah and Ning Bi.
IEEE Winter Conference on Applications of Computer Vision (WACV), 2019
Appendix Poster

DNQ: Dynamic Network Quantization
with Yuhui Xu, Yingyong Qi, Jiaxian Guo, Weiyao Lin and Hongkai Xiong.
Data Compression Conference (DCC), 2019

Understanding Straight-Through Estimator in Training Activation Quantized Neural Nets
with Penghang Yin, Jiancheng Lyu, Stanley Osher, Yingyong Qi, Jack Xin.
Seventh International Conference on Learning Representations (ICLR), 2019

Trained Rank Pruning for Efficient Deep Neural Networks
with Yuhui Xu, Yuxi Li, Wei Wen, Botao Wang, Yingyong Qi, Yiran Chen, Weiyao Lin and Hongkai Xiong.
NeurIPS 2019 Workshop on Energy Efficient Machine Learning and Cognitive Computing (EMC2), 2019
Code

A Multistage Backward Differentiable Method for Constructing Light Convolutional Neural Networks
with Fanghui Xue, Jack Xin, Jiancheng Lyu and Yingyong Qi.
International Conference on Artificial Intelligence for Industries (AI4I), 2019

Channel Pruning for Deep Neural Networks via a Relaxed Groupwise Splitting Method
with Biao Yang, Jack Xin, Jiancheng Lyu and Yingyong Qi.
International Conference on Artificial Intelligence for Industries (AI4I), 2019

TRP: Trained Rank Pruning for Efficient Deep Neural Networks
with Yuhui Xu, Yuxi Li, Wei Wen, Botao Wang, Yingyong Qi, Yiran Chen, Weiyao Lin and Hongkai Xiong.
International Joint Conference on Artificial Intelligence (IJCAI), 2020.

AutoShuffleNet: Learning Permutation Matrices via an Exact Lipschitz Continuous Penalty in Deep Convolutional Neural Networks
with Jiancheng Lyu, Yingyong Qi and Jack Xin.
26th SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2020.
one of the Most Influential KDD Papers

Edge computing
with Ying Chen and Yang Zhou.
US Patent App. 16/204,242 2020

AFIDAF: Alternating Fourier and Image Domain Adaptive Filters as an Efficient Alternative to Attention in ViTs
with Yunling Zheng, Zeyi Xu, Fanghui Xue, Biao Yang, Jiancheng Lyu, Yingyong Qi and Jack Xin.
19th International Symposium on Visual Computing (ISVC), 2024.

Applications of Computer Vision

Tiny-Hourglassnet: An Efficient Design For 3d Human Pose Estimation
with Bowen Shi, Yuhui Xu, Wenrui Dai, Botao Wang, Chenglin Li, Junni Zou and Hongkai Xiong.
2020 IEEE International Conference on Image Processing (ICIP), 2020

Weakly-Supervised Degree of Eye-Closeness Estimation
with Eyasu Mequanint, Bijan Forutanpour, Yingyong Qi and Ning Bi.
Proceedings of the IEEE International Conference on Computer Vision Workshops, 2019

User Adaptation for Biometric Authentication
with Eyasu Mequanint, Yingyong Qi and Ning Bi.
US Patent App. 16/125,360 2020

Personalized Eye Openness Estimation
with Eyasu Mequanint, Yingyong Qi and Ning Bi.
US Patent App. 16/239,352 2020

System and Method for Performing Semantic Image Segmentation
with Xiaowen Ying, Jiancheng Lyu and Yingyong Qi.
US Patent App. 17/669,040 2023

Scene segmentation and object tracking
with Jiancheng Lyu, Dashan Gao, Yingyong Qi and Ning Bi.
US Patent App. 17/828,962 2023

MobileInst: Video Instance Segmentation on the Mobile
with Renhong Zhang, Tianheng Cheng, Shusheng Yang, Haoyi Jiang, Jiancheng Lyu, Xin Li, Xiaowen Ying, Dashan Gao, Wenyu Liu and Xinggang Wang.
Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2024.

Non-convex Optimizations

Improving Network Slimming With Nonconvex Regularization
with Kevin Bui, Fredrick Park, Yingyong Qi and Jack Xin.
IEEE Access, 2021

Nonconvex Regularization for Network Slimming: Compressing CNNs Even More
with Kevin Bui, Fredrick Park, Jack Xin and Yingyong Qi.
Awarded Springer-Verlag Best Paper
15th International Symposium on Visual Computing, Oct. 5-7, 2020
Code

Structured Sparsity of Convoluational Neural Networks via Nonconvex Sparse Group Regularization
with Kevin Bui, Fredrick Park, Jack Xin and Yingyong Qi.
Frontiers in Applied Mathematics and Statistics: Mathematics of Computation and Data Science, 2020.
Code

Minimization of transformed L1 penalty: theory, difference of convex function algorithm, and robust application in compressed sensing
with Jack Xin.
Mathematical Programming, Series B, 2018
Code

Minimization of Transformed L1 Penalty: Closed Form Representation and Iterative Thresholding Algorithms
with Jack Xin.
Communications in Mathematical Sciences, 2017

Transformed Schatten-1 Iterative Thresholding Algorithms for Low Rank Matrix Completion
with Penghang Yin and Jack Xin.
Communications in Mathematical Sciences, 2017
Code

Numerical PDEs

An optimal-order error estimate for the mass-conservative characteristic finite element scheme
with Hongmei Wang and Hongxing Rui.
Applied Mathematics and Computation, 2012