Hi! I'm currently a senior Deep Learning Research Engineer at NVIDIA. I received my Master degree in Computer Vision (MSCV) from Carnegie Mellon University. Before my study at CMU, I was a Machine Learning/Computer Vision Engineer at Clobotics. I obtained my B.S. in Information Security and M.Eng. in Information and Communication Engineering from Shanghai Jiao Tong University, China.
My interests broadly lie in Computer Vision, Machine Learning and related applications. Currently my research focuses on neural network efficient training and inference, with pruning, distillation and quantization.
Publication
- Structural Pruning via Latency-Saliency Knapsack Maying Shen, Hongxu Yin, Pavlo Molchanov, Lei Mao, Jianna Liu, Jose M. Alvarez
Conference on Neural Information Processing Systems (NeurIPS), 2022
[paper] [project page] [code]
- Soft Masking for Cost-Constrained Channel Pruning Ryan Humble, Maying Shen, Jorge Albericio Latorre, Eric Darve, Jose M. Alvarez
European Conference on Computer Vision (ECCV), 2022
[paper] [code]
- Augmenting Legacy Networks for Flexible Inference Jason Clemons, Iuri Frosio, Maying Shen, Jose M. Alvarez, Stephen Keckler
European Conference on Computer Vision Workshops (ECCV/CADL), 2022
- When to Prune? A Policy towards Early Structural Pruning Maying Shen, Pavlo Molchanov, Hongxu Yin, Jose M. Alvarez
Proceedings of the IEEE/CVF International Conference on Computer Vision and Pattern Recognition (CVPR), 2022
[paper]
- Optimal Quantization Using Scaled Codebook Yerlan Idelbayev, Pavlo Molchanov, Maying Shen, Hongxu Yin, Miguel A. Carreira-Perpinan, Jose M. Alvarez
Proceedings of the IEEE/CVF International Conference on Computer Vision and Pattern Recognition (CVPR), 2021
[paper]
Education
- 2018.08 - 2019.12, Carnegie Mellon University, Robotics Insitute, School of Computer Science Master of Science in Computer Vision
Current QPA: 4.17/4.33
Courses: Introduction to Machine Learning; Computer Vision; Math Fundamentals for Robotics; Visual Learning and Recognition; Robot Localization and Mapping
- 2015.09 - 2018.03, Shanghai Jiao Tong University, Graduate School, School of Electronic Information and Electrical Engineering Master in Information and Communication Engineering
Overrall GPA: 4.11/4.3
Courses: Graph and Networks; Matrix Theory; Digital Image Processing; Natural Language Processing
- 2011.09 - 2015.06, Shanghai Jiao Tong University, School of Electronic Information and Electrical Engineering Bachelor in Information Security
Overall GPA: 88.53/100
Courses: Object Oriented Programming; Data Structure and Algorithms; Software Engineering; Computer Organization and Architecture; Course Design on Application Software; Development practice for system software
Experience
2020.02 - present: NVIDIA Corporation, Santa Clara, CA
Senior Deep Learning Research Engineer
Deep Learning Research & Development Engineer
Work on full-stack efficient deep learning.
2019.05 - 2019.08: NVIDIA Corporation, Santa Clara, CA
Deep Learning Research Intern
Worked with an AI applied researcher’s team and responsible for developing efficient and well-performed models for autonomous vehicles, in Pytorch, improving the efficiency of model inference as well as training by pruning the neural network during the training process.
2018.04 - 2018.07 / 2017.10 - 2018.01: Clobotics Co., Ltd., Shanghai, China
Machine Learning and Computer Vision Enginner
Performed large scale detection and classification of thousands of stock keeping units, in Tensorflow, achieved mid 90% precision/recall for object detection and increased coverage of stock keeping units by mid double digits percent that reached high precision/recall needed for production service level agreement.
Investigated and implemented strategy of active learning and image retrieval to improve the efficiency of data labeling.
Designed and developed modules using Python and OpenCV to automatically process turn-table taken images to increase the volume and diversity of training data, to optimize the fine-grained classification with minor size and appearance variation.
Machine Learning Intern
Designed and implemented algorithm of image stitching for images taken from multiple camera viewpoints and angles with potentially hostile lighting condition to generate a large panoramic image.
Founding and core contributor of the real time multiple images object detection service with deep learning models and the stitching algorithm.
2015.09 - 2017.06: Information Cognition Lab, Shanghai Jiao Tong University, Shanghai, China
Graduate Research Assistant
Worked with Prof. Xinghao Jiang and Dr. Tanfeng Sun, mainly focused on crowd counting estimation and anomaly detection.
2014.07 - 2015.02: Intel Asia-Pacific Research & Development Ltd., Shanghai, China
Software Engineering Intern
Developed distributed framework in Python to have jobs run automatically to test processors, with functions of auto-trigger, job assignment, output files collection, report generation and notification sent by email.
Contributed to developing encryption module based on AES for android platform attendance system.