About

Dr. Chen Shangyu

Algorithm Engineer in Alibaba, Ph.D student from Nanyang Technological University
School of Computer Science and Engineering, Nanyang Technological University
Under supervision of Prof. Sinno Jialin Pan

Email: schen025 [at] e.ntu.edu.sg

CV: CV

Research Interests

Deep Neural Network Model Compression (Pruning, Quantization)

Education Experience

2012-2016 BS

School of Data and Computer Science, Sun Yat-Sen University, Guangzhou, China

2016-Now Ph.D

School of Computer Science and Engineering, Nanyang Technological University, Singapore

Career Experience

April, 2020 - Now Algorithm Engineer II

Search Rank Team (Taobao), Alibaba, Hangzhou, China.

June-Nov, 2020 Research Intern

ByteDance AI Lab, Language and Information Technology Group in Singapore. Honor to work with Dr. Li Hang

Publication

• Jianda Chen, Shangyu Chen and Sinno Jialin Pan. Storage Efficient and Dynamic Flexible Runtime Channel Pruning via Deep Reinforcement Learning. To be appear in 34rd Conference on Neural Information Processing Systems 2020 (NeurIPS-20). Virtual. Dec. 6-12, 2020. (Poster)
• Shangyu Chen, Wenya Wang and Sinno Jialin Pan. MetaQuant: Learning to Quantize by Learning to Penetrate Non-differentiable Quantization. Published in Proceedings of the 33rd Conference on Neural Information Processing Systems 2019 (NeurIPS-19). Vancouver, Canada. Dec. 8-14, 2019. (Poster) [code] [paper]
• Shangyu Chen, Wenya Wang and Sinno Jialin Pan. Cooperative Pruning in Cross-Domain Deep Neural Network Compression. Published in Proceedings of the 28th International Joint Conference on Artificial Intelligence (IJCAI-19). Macao, China. Aug. 10 - 16, 2019. (Oral) [code] [paper]
• Shangyu Chen, Wenya Wang and Sinno Jialin Pan. Deep Neural Network Quantization via Layer-Wise Optimization using Limited Training Data. Published in Proceedings of the 33rd AAAI Conference on Artificial Intelligence (AAAI-19). Hawaii, USA. Jan. 27 - Feb. 1, 2019. (Oral) [code][paper]
• Xin Dong, Shangyu Chen and Sinno Jialin Pan. Learning to Prune Deep Neural Networks via Layer-wise Optimal Brain Surgeon To appear in Annual Conference on Neural Information Processing Systems 2017 (NIPS-17). Long Beach, USA. Dec. 4-9, 2017. (Poster) [code] [paper]

Professional Services

Conference Reviewers

• International Conference on Machine Learning: 2019/2020/2021
• Conference on Neural Information Processing Systems (NeurIPS 2019) (Top 50% best reviewers), 2020/2021
• AAAI Conference on Artificial Intelligence (AAAI 2020/2021)
• International Conference on Learning Representations (ICLR 2021)

Research Projects

• Awesome Deep Neural Network Compression
  • Listing and Summary of related papers.
  • Re-implementation of state-of-the-art methods.

Teaching Assistant

• CX/CZ 1007 Data Strcuture
• CE/CZ 4041 Data Mining
• CZ 3005 Artificial Intelligence

Applied Project

Undergraduate Period

1. Image Recommendation System

• Performed image recommendation based on images’ SIFT-saliency and user browsing history.

• Project Homepage.
  

2. Lane Detection and Inverse Perspective Mapping Generation

• Designed and implemented an innovative algorithm to perform automatic Inverse Perspective Mapping of road, which tradditionally required camera to be calibrated maunally in advance.

• Performed lane detection based on inverse perspective mapping images.

• Project Homepage