Article
Computer Science, Hardware & Architecture
Surong Yan, Haosen Wang, Yixiao Li, Chunqi Wu, Long Han, Chenglong Shi, Ruilin Guo
Summary: MFGRec is a recommendation method based on heterogeneous information network, which leverages the semantic and structural features of metapath to improve recommendation performance. The model filters a large amount of noise and irrelevant information from intra-metapath and inter-metapath perspectives, significantly improving the scalability and accuracy of the recommendation framework.
JOURNAL OF SUPERCOMPUTING
(2023)
Article
Mathematics
Peisen Yuan, Yi Sun, Hengliang Wang
Summary: Recommendation systems are widely used on the internet to predict user preferences through interactions with products, utilizing heterogeneous information networks to address data challenges and search for metapaths suitable for different recommendation tasks.
Article
Computer Science, Information Systems
Ying Ji, Guojia Wan, Yibing Zhan, Bo Du
Summary: In this paper, we propose a method to model a molecule as a heterogeneous graph and leverage metapaths to capture latent features for chemical functional groups. We construct metapath-based connectivity and decompose the heterogeneous graph into subgraphs according to relation types. A hierarchical attention strategy is designed to aggregate heterogeneous information at the node and relation level. Experimental results show the effectiveness of our model with competitive performance.
INFORMATION SCIENCES
(2023)
Article
Mathematics
Chengdong Zhang, Keke Li, Shaoqing Wang, Bin Zhou, Lei Wang, Fuzhen Sun
Summary: In this paper, a novel definition of a metapath is proposed, which integrates the edge type into it. The embedding of nodes is trained by encoding and aggregating the neighbors of intrapaths, and the final embedding is obtained by aggregating nodes from interpaths using attention mechanism. Experimental results show that the proposed method outperforms the state-of-the-art baselines in link prediction tasks.
Article
Computer Science, Artificial Intelligence
Mengya Guan, Xinjun Cai, Jiaxing Shang, Fei Hao, Dajiang Liu, Xianlong Jiao, Wancheng Ni
Summary: This paper proposes a new HGNN model named HMSG, which can comprehensively capture structural, semantic, and attribute information from both homogeneous and heterogeneous neighbors more purposefully. The performance of HMSG is evaluated through multiple graph-mining tasks and outperforms state-of-the-art baselines.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Theory & Methods
Jongmin Park, Soohwan Jeong, Byung Suk Lee, Sungsu Lim
Summary: This paper proposes a new model called MIGTNet for heterogeneous graph embedding, which uses both metapath instances and relations between them. MIGTNet constructs a metapath instance-based graph, where a node represents a metapath instance and a link represents a relation between metapath instances, and inputs it to a hierarchical graph attention network to obtain meaningful node embeddings. Extensive experiments show that MIGTNet outperforms state-of-the-art heterogeneous graph embedding models in node classification and node clustering.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Biochemical Research Methods
Mei Li, Xiangrui Cai, Sihan Xu, Hua Ji
Summary: This paper proposes a metapath-aggregated heterogeneous graph neural network (MHGNN) for drug-target interaction (DTI) prediction. MHGNN can capture complex structures and rich semantics in the biological heterogeneous graph, and achieves favorable results in DTI prediction.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Chemistry, Multidisciplinary
Xiaole Wang, Jiwei Qin, Shangju Deng, Wei Zeng
Summary: This paper proposes a knowledge-aware enhanced network (KCNR) to address the insufficient representation of user and item embeddings in knowledge graph-based recommendation methods. By propagating neighborhood information in the knowledge graph, KCNR enriches user descriptions and enhances item embedding representations. Furthermore, KCNR utilizes an information complementarity module to share potential interaction characteristics, automatically discovering semantic information and capturing users' personalized preferences. Extensive experiments in movies, books, and music demonstrate the excellent recommendation performance of KCNR.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Yaomin Chang, Lin Shu, Erxin Du, Chuan Chen, Ziyang Zhang, Zibin Zheng, Yuzhao Huang, Xingxing Xing
Summary: Reciprocal Recommender Systems (RRSs) are recommender systems designed for people-to-people recommendation tasks. In this paper, a novel Graph neural network for Reciprocal Recommendation (GraphRR) is proposed to utilize users' multiplex interactions. Experimental results demonstrate the superiority of GraphRR and provide empirical evidence for the benefits of the proposed ego graph augmentation.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Xiaoru Chen, Yingxu Wang, Jinyuan Fang, Zaiqiao Meng, Shangsong Liang
Summary: Heterogeneous Graph Contrastive learning model with Metapath-based Augmentations (HGCMA) is proposed in this paper to learn discriminative representations of nodes in a heterogeneous graph. By utilizing metapaths to construct graphs, perform data augmentation and optimize the contrastive objective, HGCMA achieves effective node embeddings and outperforms state-of-the-art methods, as validated by extensive experiments.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Biochemical Research Methods
Mei Li, Xiangrui Cai, Sihan Xu, Hua Ji
Summary: In this paper, the authors propose a metapath-aggregated heterogeneous graph neural network (MHGNN) for drug-target interaction (DTI) prediction. By modeling high-order relations via metapaths, MHGNN is able to capture complex structures and rich semantics in the biological heterogeneous graph. Experimental results demonstrate that MHGNN outperforms 17 state-of-the-art methods in drug repositioning.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Tianjun Wei, Tommy W. S. Chow
Summary: This paper introduces a unified graph recommendation framework that combines graph convolution networks (GCN) and traditional models, and proposes a novel Fused Graph Context-aware Recommender system (FGCR) model to address limitations of existing models. FGCR models robust node relationships in the user-item-context interaction graph using a sparse item correlation matrix and dense node embeddings, and applies a novel masked graph convolution strategy for refining the information aggregation process. Experimental results show that FGCR significantly outperforms seven baseline models. Ablation study and user group analysis validate the effectiveness of each component in FGCR, particularly in modeling active users and sparse contextual information.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Zhonghai He, Bei Hui, Shengming Zhang, Chunjing Xiao, Ting Zhong, Fan Zhou
Summary: Knowledge graph-based recommendation models use auxiliary information to address sparsity and cold-start problems. However, existing approaches neglect indirect feedback and diversity of multi-hop neighbors. We propose a novel recommender system that enhances representation learning by considering these factors.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Surong Yan, Chongyang Li, Haosen Wang, Bin Lin, Yixian Yuan
Summary: This paper introduces a Feature Interactive Graph Neural Network for KG-based Recommendation (FIKGRec) to improve the performance of recommendation. The method models the interactions between nodes in the knowledge graph and designs a preference-aware attention mechanism to capture the user's fine-grained preference. Experiments demonstrate that the proposed method outperforms existing methods.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Computer Science, Artificial Intelligence
Di Wu, Mingjing Tang, Shu Zhang, Ao You, Wei Gao
Summary: This paper proposes a deep knowledge preference-aware reinforcement learning network (KPRLN) for recommendation, which builds paths between user's historical interaction items in the knowledge graph, learns the preference features of each user-entity-relation, and generates the weighted knowledge graph with fine-grained preference features. Extensive experiments on two real-world datasets demonstrate that our method outperforms other state-of-the-art baselines.
COMPLEX & INTELLIGENT SYSTEMS
(2023)
Article
Development Studies
Yizhong Huan, Lingqing Wang, Mark Burgman, Haitao Li, Yurong Yu, Jianpeng Zhang, Tao Liang
Summary: The 17 Sustainable Development Goals (SDGs) proposed in 2015 require significant costs to achieve, and a prioritization framework is needed to reduce complexity and lower costs. This study creates a new composite assessment framework by synthesizing multiple methods, and tests its effectiveness by ranking prioritizations for SDG 6 targets in Southeast Asia.
SUSTAINABLE DEVELOPMENT
(2022)
Article
Computer Science, Information Systems
Feng Zhao, Xiangyu Gui, Yafan Huang, Hai Jin, Laurence T. Yang
IEEE TRANSACTIONS ON BIG DATA
(2020)
Proceedings Paper
Computer Science, Software Engineering
Bohan Zhang, Yafan Huang, Guanpeng Li
Summary: This paper proposes Salus, a data-driven real-time safety monitor that detects and mitigates safety violations of autonomous vehicles. By using machine learning techniques and learning from the safety violations of the vehicles, Salus models traffic behaviors, characterizes early symptoms, and deploys real-time safety violation detection before actual crashes happen. The evaluation demonstrates that the proposed technique is highly effective in reducing safety violations in industry-level autonomous driving systems.
2022 IEEE 22ND INTERNATIONAL CONFERENCE ON SOFTWARE QUALITY, RELIABILITY AND SECURITY, QRS
(2022)
Proceedings Paper
Computer Science, Software Engineering
Bohan Zhang, Yafan Huang, Rachael Chen, Guanpeng Li
Summary: This paper proposes D2MON, a data-driven real-time safety monitor, for detecting and mitigating safety violations of autonomous vehicles. By learning from existing safety violations, the system can identify traffic situations that lead to safety violations and detect their symptoms in advance. It takes safety actions if dangerous surroundings are detected, ensuring the AV remains safe in the evolving traffic environment.
2022 IEEE INTERNATIONAL SYMPOSIUM ON SOFTWARE RELIABILITY ENGINEERING WORKSHOPS (ISSREW 2022)
(2022)
Proceedings Paper
Computer Science, Software Engineering
Yafan Huang, Shengjian Guo, Sheng Di, Guanpeng Li, Franck Cappello
Summary: The study finds that the existing SID technique faces a decrease in SDC coverage in HPC applications, due to evaluation limitations to single program inputs. To address this issue, the Sentinel framework is proposed to enhance SDC coverage across multiple inputs through automated compiler techniques.
PPOPP'22: PROCEEDINGS OF THE 27TH ACM SIGPLAN SYMPOSIUM ON PRINCIPLES AND PRACTICE OF PARALLEL PROGRAMMING
(2022)
Proceedings Paper
Computer Science, Cybernetics
Shihui Song, Yafan Huang, Hongwei Lu
Summary: With the growing popularity of social media and the challenges posed by rumor propagation, detecting rumors accurately and handling imbalanced data are critical. The OID-GCN model, which integrates Katz centrality with spectral-domain graph convolution, effectively addresses these issues and outperforms existing methods in comprehensive experimental results on real-world datasets.
2021 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS (SMC)
(2021)