Jun Zhao

Jun Zhao

Auto-generated profile, based on publicly available information

Jun Zhao

Kuaishou, Beijing, China

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
21 October 2023
Verified
Conference Paper
Open access
BZBeichuan Zhang
Beichuan Zhang
CSChenggen Sun
Chenggen Sun
JTJianchao Tan
Jianchao Tan
XCXinjun Cai
Xinjun Cai
JZJun Zhao
Jun Zhao
YSYang Song
Yang Song
Increasing the size of embedding layers has shown to be effective in improving the performance of recommendation models, yet gradually causing their sizes to exceed terabytes in industrial recommender systems, and hence the increase of computing and storage costs. To save resources while maintaining model performances, we propose SHARK, the model compression practice we have summarized in the recommender system of industrial scenarios. SHARK consists of two main components. First, we use the novel first-order component of Taylor expansion as importance scores to prune the number of embedding tables (feature fields). Second, we introduce a new row-wise quantization method to apply different quantization strategies to each embedding. We conduct extensive experiments on both public and industrial datasets, demonstrating that each component of our proposed SHARK framework outperforms previous approaches. We conduct A/B tests in multiple models on Kuaishou, such as short video, e-commerce, and advertising recommendation models. The results of the online A/B test showed SHARK can effectively reduce the memory footprint of the embedded layer. For the short-video scenarios, the compressed model without any performance drop significantly saves 70% storage and thousands of machines, improves 30% queries per second (QPS), and has been deployed to serve hundreds of millions of users and process tens of billions of requests every day.

2 October 2019
Verified
Book Chapter
Closed access
PDF file
CLChengfan Li
Chengfan Li
LLLan Liu
Lan Liu
XSXian Kun Sun
Xian Kun Sun
JZJun Juan Zhao
Jun Juan Zhao
YYYuan Yuan
Yuan Yuan
JYJing Yuan Yin
Jing Yuan Yin

16 June 2023
Verified
Conference Paper
Closed access
PDF file
RXRongji Xie
Rongji Xie
JZJun Zhao
Jun Zhao
Xu ChenXu Chen
Xu Chen
ZCZheng Chen
Zheng Chen
Most of the sound absorption methods in noise control used fibrous materials, which are prone to generate dust particles and other problems. Micro-perforated sound absorption structure can have good sound absorption performance without filling any sound absorption materials. It can effectively solve the problem of dust particles caused by fibrous soundabsorbing materials. Based on the micro-perforation theory, the sound absorption structure of double-layer composite metal plate is studied according to the acoustic and electrical analogy method, and the sound absorption performance of different size structures is calculated and simulated in MATLAB. The two structures with ideal simulation results are processed for sample preparation, and the performance test verification is completed in the reverberation room.