Weijin Liu

Weijin Liu

Auto-generated profile, based on publicly available information

Weijin Liu

School of Materials Science and Engineering, South China University of Technology, Guangzhou, China

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

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

Representative Publications

View All Publications

19 July 2019 in Fibers and Polymers
Verified
Journal Article
Closed access
PDF file
JSJieyao Song
Jieyao Song
PCPengran Chen
Pengran Chen
WLWeijin Liu
Weijin Liu

9 July 2019 in Journal of Dispersion Science and Technology
Verified
Journal Article
Closed access
MXMimi Xiong
Mimi Xiong
ZRZhonghai Ren
Zhonghai Ren
WLWeijin Liu
Weijin Liu
A hybrid UV absorber (SiO2-H) was successfully synthesized by functionalizing nanosilica (SiO2) with 2,4-dihydroxybenzophenone and then was introduced onto the surface of cotton fabric by sequentially depositing in poly (diallyl dimethyl ammonium chloride) (PDDA) aqueous solution and SiO2-H suspension via layer-by-layer assembly method. The SiO2-H particles imparted excellent ultraviolet (UV) resistance property to the cotton fabric, while forming nano- and micro structure on the surface of cotton fabric, making the fabric superhydrophobic after the modification with hexadecyltrimethoxy silane (HDTMS). The surface morphology, wetting behavior, UV resistance and physical properties of cotton fabrics were studied. According to the results, the modified cotton fabric not only showed strong UV-resistant ability with Ultraviolet Protection Factor (UPF) value as high as 500, but also displayed superhydrophobicty with WCA of 153 ± 1°. Moreover, the durability of multifunctional surface was examined by laundering and abrasion tests. The modified fabrics still maintained hydrophobic property and excellent UV resistance after 300 abrasion or 20 laundering cycles, demonstrating outstanding durability.

21 July 2022 in Journal of Materials Science: Materials in Electronics
Verified
Journal Article
Closed access
PDF file
PGPeixin Gao
Peixin Gao
WLWeijin Liu
Weijin Liu
JYJin Yang
Jin Yang

28 February 2025
Verified
Conference Paper
Open access
BSBaixi Sun
Baixi Sun
WLWeijin Liu
Weijin Liu
JPJ. Gregory Pauloski
J. Gregory Pauloski
JTJiannan Tian
Jiannan Tian
JJJinda Jia
Jinda Jia
DTDingwen Tao
Dingwen Tao
Second-order optimization methods have been developed to enhance convergence and generalization in deep neural network (DNN) training compared to first-order methods like Stochastic Gradient Descent (SGD). However, these methods face challenges in distributed settings due to high communication overhead. Gradient compression, a technique commonly used to accelerate communication for first-order approaches, often results in low communication reduction ratios, decreased model accuracy, and/or high compression overhead when applied to second-order methods. To address these limitations, we introduce a novel gradient compression method for second-order optimizers called COMPSO. This method effectively reduces communication costs while preserving the advantages of second-order optimization. COMPSO employs stochastic rounding to maintain accuracy and filters out minor gradients to improve compression ratios. Additionally, we develop GPU optimizations to minimize compression overhead and performance modeling to ensure end-to-end performance gains across various systems. Evaluation of COMPSO on different DNN models shows that it achieves a compression ratio of 22.1×, reduces communication time by 14.2×, and improves overall performance by 1.9×, all without any drop in model accuracy.

24 August 2014 in Journal of Thermal Analysis and Calorimetry
Verified
Journal Article
Closed access
PDF file
JCJunzhao Chen
Junzhao Chen
XWXiaobin Wang
Xiaobin Wang
WLWeijin Liu
Weijin Liu