Search

Your search keyword '"Liang, Qiuyu"' showing total 20 results
Start Over Author "Liang, Qiuyu"
20 results on '"Liang, Qiuyu"'

1. Fully Hyperbolic Rotation for Knowledge Graph Embedding

Author

Liang, Qiuyu, Wang, Weihua, Bao, Feilong, and Gao, Guanglai

Subjects
Computer Science - Artificial Intelligence, Computer Science - Machine Learning
Abstract

Hyperbolic rotation is commonly used to effectively model knowledge graphs and their inherent hierarchies. However, existing hyperbolic rotation models rely on logarithmic and exponential mappings for feature transformation. These models only project data features into hyperbolic space for rotation, limiting their ability to fully exploit the hyperbolic space. To address this problem, we propose a novel fully hyperbolic model designed for knowledge graph embedding. Instead of feature mappings, we define the model directly in hyperbolic space with the Lorentz model. Our model considers each relation in knowledge graphs as a Lorentz rotation from the head entity to the tail entity. We adopt the Lorentzian version distance as the scoring function for measuring the plausibility of triplets. Extensive results on standard knowledge graph completion benchmarks demonstrated that our model achieves competitive results with fewer parameters. In addition, our model get the state-of-the-art performance on datasets of CoDEx-s and CoDEx-m, which are more diverse and challenging than before. Our code is available at https://github.com/llqy123/FHRE., Comment: Accepted by ECAI 2024

Published
2024
Full Text
View/download PDF

2. L^2GC:Lorentzian Linear Graph Convolutional Networks for Node Classification

Author

Liang, Qiuyu, Wang, Weihua, Bao, Feilong, and Gao, Guanglai

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Computation and Language
Abstract

Linear Graph Convolutional Networks (GCNs) are used to classify the node in the graph data. However, we note that most existing linear GCN models perform neural network operations in Euclidean space, which do not explicitly capture the tree-like hierarchical structure exhibited in real-world datasets that modeled as graphs. In this paper, we attempt to introduce hyperbolic space into linear GCN and propose a novel framework for Lorentzian linear GCN. Specifically, we map the learned features of graph nodes into hyperbolic space, and then perform a Lorentzian linear feature transformation to capture the underlying tree-like structure of data. Experimental results on standard citation networks datasets with semi-supervised learning show that our approach yields new state-of-the-art results of accuracy 74.7$\%$ on Citeseer and 81.3$\%$ on PubMed datasets. Furthermore, we observe that our approach can be trained up to two orders of magnitude faster than other nonlinear GCN models on PubMed dataset. Our code is publicly available at https://github.com/llqy123/LLGC-master., Comment: Accepted by LREC-COLING 2024

Published
2024

11. The role of reduced graphene oxide as a '2D flexible crosslinking point' in composite hydrogels

Author

Xia, Shuang, Feng, Shi, Deng, Zejiang, Liang, Qiuyu, Xiang, Xu, and Xie, Hongyan

Subjects
Composite materials -- Materials -- Mechanical properties, Graphene -- Usage, Engineering and manufacturing industries, Science and technology
Abstract

The composite hydrogels are widely used in various research directions. How to improve the mechanical properties is one of the concerns of researchers. In this work, the mechanical properties of composite hydrogels with reduced graphene oxide (rGO), kaolin (KL), montmorillonite (MMT), and attapulgite (APT) as additives were compared. The morphology of polyacrylamide (PAM) composite hydrogel was observed by electron microscope, and the hydrogen bonding between inorganic additives and PAM chain in the hydrated samples were detected by IR. The rheological data (G, G', and tan [delta]) confirmed the beneficial effect of rGO as a two-dimension (2D) material on the mechanical properties of hydrogels. In the tensile-fracture experiments, the maximum tensile strain of PAM-rGO was 568% and the maximum tensile force was 1.18 N, both of which were larger than that of the clay-based composite hydrogel with the same mass concentration. Based on these experimental data, the response behavior of rGO in the process of hydrogel deformation was elucidated by the concept of '2D flexible crosslinking point,' and the reason why PAM-rGO composite hydrogel had better mechanical strength was explained in detail. KEYWORDS 2D flexible crosslinking point, composite hydrogel, mechanical properties, reduced graphene oxide, two-dimensional materials, 1 | INTRODUCTION Hydrogels are materials composed of water and polymer matrix. Because of its excellent properties, especially good biocompatibility, hydrogels have attracted more and more attention in the fields [...]

Published
2022
Full Text
View/download PDF

14. L$^2$GC: Lorentzian Linear Graph Convolutional Networks For Node Classification

Author

Liang, Qiuyu, Wang, Weihua, Bao, Feilong, Gao, Guanglai, Liang, Qiuyu, Wang, Weihua, Bao, Feilong, and Gao, Guanglai

Abstract

Linear Graph Convolutional Networks (GCNs) are used to classify the node in the graph data. However, we note that most existing linear GCN models perform neural network operations in Euclidean space, which do not explicitly capture the tree-like hierarchical structure exhibited in real-world datasets that modeled as graphs. In this paper, we attempt to introduce hyperbolic space into linear GCN and propose a novel framework for Lorentzian linear GCN. Specifically, we map the learned features of graph nodes into hyperbolic space, and then perform a Lorentzian linear feature transformation to capture the underlying tree-like structure of data. Experimental results on standard citation networks datasets with semi-supervised learning show that our approach yields new state-of-the-art results of accuracy 74.7$\%$ on Citeseer and 81.3$\%$ on PubMed datasets. Furthermore, we observe that our approach can be trained up to two orders of magnitude faster than other nonlinear GCN models on PubMed dataset. Our code is publicly available at https://github.com/llqy123/LLGC-master., Comment: Accepted by LREC-COLING 2024

Published
2024

15. A Simple, Rapid, and Green Method for Preparing Highly Conductive Reduced Graphene Oxide Aerogel for Pressure Sensors.

Author

Zhang, Haifeng, Feng, Shi, Yang, Jing, Liang, Qiuyu, and Xiang, Xu

Abstract

Due to its excellent performance, reduced graphene oxide aerogel (rGA) has been widely used in recent years. Typically, the mechanical and electrical properties of graphene oxide can be enhanced through doping, compositing, or heat treatment. This work presents a simple, fast, and environmentally friendly method for producing high-performance rGA using two reducing agents. During the reduction process, ascorbic acid can cross-link and reduce graphene oxide, while the reducing ability of ammonium citrate restores the conjugated structure of the graphene sheets. By fixing the concentration of ascorbic acid and adjusting the amount of ammonium citrate, we can obtain DrGA-2 that possesses both mechanical and electrical properties. The maximum strain and stress of DrGA-2 are 40.0% and 21.5 kPa, respectively. Meanwhile, only a 10% compressive strain is sufficient to induce a 95% resistance change rate. When DrGA-2 is assembled into a capacitive sensor, its capacitance change rate ranges from 80% to 200% under different strain conditions, enabling real-time monitoring of subtle human movements. This demonstrates the tremendous potential of this material in this field. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

16. Effects of NAD+ precursors on blood pressure, C‐reactive protein concentration and carotid intima‐media thickness: A meta‐analysis of randomized controlled trials

Author

Lei, Langhuan, primary, Zhang, Xiaoli, additional, Lin, Jiali, additional, Liang, Qiuyu, additional, Sohouli, Mohammad Hassan, additional, Izze da Silva Magalhães, Elma, additional, Fatahi, Somaye, additional, Yang, Lihua, additional, Xu, Wanting, additional, Wang, Xingyong, additional, Li, Wei, additional, and Yang, Jianrong, additional

Published
2023
Full Text
View/download PDF

20. COMPARISON OF CHINESE AND AMERICAN CACTUS PEAR (OPUNTIA CACTI) IN INDUCTION OF ROS AND GROWTH INHIBITION IN BLADDER CANCER CELLS

Author

Liang QiuYu, F. A. P. Campos, Jean M. Feugang, Zou ChangPing, Wu Qian, Liu HuaGuang, D. Zou, Bai YaLai, José C. B. Dubeux, James Lin, and S. de Melo Silva

Subjects
PEAR, Bladder cancer, Traditional medicine, Horticulture, Biology, medicine.disease, chemistry.chemical_compound, chemistry, Cell culture, Apoptosis, Cactus, Botany, medicine, Growth inhibition, Medicinal plants, Cytotoxicity
Published
2009

Catalog

Books, media, physical & digital resources
See catalog results